{"id":33954,"date":"2025-11-09T07:41:35","date_gmt":"2025-11-09T04:41:35","guid":{"rendered":"https:\/\/fatihsoysal.com\/blog\/?p=33954"},"modified":"2025-11-09T07:41:35","modified_gmt":"2025-11-09T04:41:35","slug":"yolov12-nedir-ve-neden-onemlidir","status":"publish","type":"post","link":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/","title":{"rendered":"YOLOv12 Nedir ve Neden \u00d6nemlidir?"},"content":{"rendered":"<p><body><\/p>\n<p>Ger\u00e7ek zamanl\u0131 nesne tespiti, bilgisayar g\u00f6r\u00fc\u015f\u00fc alan\u0131n\u0131n en heyecan verici ve h\u0131zl\u0131 geli\u015fen dallar\u0131ndan biridir. Otonom ara\u00e7lardan g\u00fcvenlik sistemlerine, end\u00fcstriyel otomasyondan art\u0131r\u0131lm\u0131\u015f ger\u00e7ekli\u011fe kadar pek \u00e7ok uygulamada kritik bir rol oynamaktad\u0131r. Bu alandaki devrim niteli\u011findeki geli\u015fmelerin \u00f6nc\u00fcs\u00fc olan YOLO (You Only Look Once) serisi, h\u0131z\u0131 ve do\u011frulu\u011fu bir araya getirerek sekt\u00f6re y\u00f6n vermeye devam ediyor. Her yeni s\u00fcr\u00fcmle birlikte performans \u00e7\u0131tas\u0131n\u0131 daha da yukar\u0131 ta\u015f\u0131yan YOLO ailesi, \u015fimdi de <strong>YOLOv12<\/strong> ile &#8220;ger\u00e7ek zamanl\u0131&#8221; kavram\u0131n\u0131 yeniden tan\u0131mlamaya haz\u0131rlan\u0131yor.<\/p>\n<p>YOLOv12, \u00f6nceki s\u00fcr\u00fcmlerin miras\u0131n\u0131 devralarak, yapay zeka ve derin \u00f6\u011frenme teknolojilerindeki en son yenilikleri b\u00fcnyesinde bar\u0131nd\u0131r\u0131yor. Ama\u00e7, sadece daha h\u0131zl\u0131 veya daha do\u011fru olmakla kalmay\u0131p, ayn\u0131 zamanda daha verimli, daha esnek ve daha kolay da\u011f\u0131t\u0131labilir bir \u00e7\u00f6z\u00fcm sunmakt\u0131r. Bu teknik rehberde, YOLOv12&#8217;nin mimari yeniliklerini, performans ba\u015far\u0131lar\u0131n\u0131, pratik uygulama ad\u0131mlar\u0131n\u0131 ve gelecekteki potansiyelini derinlemesine inceleyece\u011fiz.<\/p>\n<h2>YOLOv12 Nedir ve Neden \u00d6nemlidir?<\/h2>\n<p>YOLO serisi, tek bir sinir a\u011f\u0131 ge\u00e7i\u015fiyle hem nesnelerin konumunu hem de s\u0131n\u0131flar\u0131n\u0131 tahmin ederek, nesne tespitini bir regresyon problemi olarak ele alan ilk modellerden biriydi. Bu yakla\u015f\u0131m, geleneksel iki a\u015famal\u0131 dedekt\u00f6rlere (\u00f6rne\u011fin R-CNN ailesi) k\u0131yasla devasa bir h\u0131z avantaj\u0131 sa\u011flad\u0131. YOLOv1&#8217;den ba\u015flayarak, her yeni s\u00fcr\u00fcmde mimari iyile\u015ftirmeler, daha iyi \u00f6zellik \u00e7\u0131kar\u0131m\u0131, geli\u015fmi\u015f kay\u0131p fonksiyonlar\u0131 ve daha verimli e\u011fitim stratejileri ile performans s\u00fcrekli art\u0131r\u0131ld\u0131.<\/p>\n<p><strong>YOLOv12<\/strong>, bu evrimin en son ve en geli\u015fmi\u015f halkas\u0131d\u0131r. Temel amac\u0131, \u00e7e\u015fitli donan\u0131m platformlar\u0131nda (g\u00f6m\u00fcl\u00fc cihazlardan bulut sunucular\u0131na kadar) \u00fcst\u00fcn do\u011fruluk ve inan\u0131lmaz h\u0131z sunmakt\u0131r. \u00d6nceki s\u00fcr\u00fcmlerin (YOLOv5, YOLOv7, YOLOv8, YOLOv9 vb.) g\u00fc\u00e7l\u00fc y\u00f6nlerini birle\u015ftirirken, ayn\u0131 zamanda yeni nesil derin \u00f6\u011frenme tekniklerini entegre ederek b\u00fcy\u00fck bir s\u0131\u00e7rama yapmaktad\u0131r. Bu s\u00fcr\u00fcm, \u00f6zellikle d\u00fc\u015f\u00fck gecikme s\u00fcresi gerektiren uygulamalar i\u00e7in tasarlanm\u0131\u015f olup, daha az hesaplama kayna\u011f\u0131yla daha iyi sonu\u00e7lar elde etmeyi hedefler.<\/p>\n<p>YOLOv12&#8217;nin \u00f6nemi, sadece teknik ba\u015far\u0131lar\u0131yla s\u0131n\u0131rl\u0131 de\u011fildir. End\u00fcstriyel uygulamalarda, otonom s\u00fcr\u00fc\u015fte, robotikte, sa\u011fl\u0131k hizmetlerinde ve g\u00fcvenlik sistemlerinde devrim yaratma potansiyeli ta\u015f\u0131maktad\u0131r. Daha h\u0131zl\u0131 ve daha do\u011fru tespitler, bu alanlarda daha g\u00fcvenli, daha verimli ve daha ak\u0131ll\u0131 sistemlerin geli\u015ftirilmesine olanak tan\u0131r. YOLOv12, geli\u015ftiricilere ve ara\u015ft\u0131rmac\u0131lara, ger\u00e7ek d\u00fcnya problemlerini \u00e7\u00f6zmek i\u00e7in g\u00fc\u00e7l\u00fc ve eri\u015filebilir bir ara\u00e7 sunarak, bilgisayar g\u00f6r\u00fc\u015f\u00fcn\u00fcn s\u0131n\u0131rlar\u0131n\u0131 zorlamaktad\u0131r.<\/p>\n<h2>YOLOv12&#8217;nin Mimari Yenilikleri ve Temel Bile\u015fenleri<\/h2>\n<p>YOLOv12&#8217;nin kalbinde, performans\u0131 maksimize etmek i\u00e7in tasarlanm\u0131\u015f sofistike bir mimari yat\u0131yor. Bu mimari, genellikle \u00fc\u00e7 ana b\u00f6l\u00fcmden olu\u015fur: \u00d6zellik \u00c7\u0131kar\u0131c\u0131 (Backbone), Yol Toplay\u0131c\u0131 (Neck) ve Tespit Kafas\u0131 (Head). YOLOv12, bu b\u00f6l\u00fcmlerin her birinde \u00f6nemli yenilikler sunmaktad\u0131r.<\/p>\n<h3>\u00d6zellik \u00c7\u0131kar\u0131c\u0131 (Backbone): Dinamik Mod\u00fcler A\u011f (DMA)<\/h3>\n<p>YOLOv12&#8217;nin omurgas\u0131, y\u00fcksek verimlilik ve g\u00fc\u00e7l\u00fc \u00f6zellik temsili i\u00e7in tasarlanm\u0131\u015f Dinamik Mod\u00fcler A\u011f (DMA) ad\u0131n\u0131 verdi\u011fimiz yeni nesil bir yap\u0131ya sahiptir. Bu backbone, geleneksel evri\u015fimsel katmanlar\u0131, daha az parametreyle daha geni\u015f bir al\u0131c\u0131 alan (receptive field) sa\u011flayan seyreltilmi\u015f (dilated) evri\u015fimler ve hesaplama maliyetini d\u00fc\u015f\u00fcren derinlemesine ayr\u0131labilir (depthwise separable) evri\u015fimler ile birle\u015ftirir. Ayr\u0131ca, dinamik aktivasyon fonksiyonlar\u0131 ve dikkat mekanizmalar\u0131 (\u00f6rne\u011fin, uzamsal ve kanal tabanl\u0131 dikkat) sayesinde, modelin giri\u015f g\u00f6r\u00fcnt\u00fcs\u00fcndeki \u00f6nemli b\u00f6lgelere daha fazla odaklanmas\u0131 sa\u011flan\u0131r. Bu, \u00f6zellikle k\u00fc\u00e7\u00fck nesnelerin tespitinde ve karma\u015f\u0131k sahnelerde \u00f6nemli bir avantaj sa\u011flar.<\/p>\n<h3>Yol Toplay\u0131c\u0131 (Neck): \u00c7ok \u00d6l\u00e7ekli Adaptif F\u00fczyon (MSAF)<\/h3>\n<p>Neck k\u0131sm\u0131, backbone&#8217;dan gelen farkl\u0131 \u00f6l\u00e7eklerdeki \u00f6zellik haritalar\u0131n\u0131 birle\u015ftirerek, tespit kafas\u0131na daha zengin ve ba\u011flamsal bilgi sa\u011flamaktan sorumludur. YOLOv12, \u00c7ok \u00d6l\u00e7ekli Adaptif F\u00fczyon (MSAF) ad\u0131 verilen geli\u015ftirilmi\u015f bir Path Aggregation Network (PANet) ve Bi-directional Feature Pyramid Network (BiFPN) hibrit yap\u0131s\u0131n\u0131 kullan\u0131r. MSAF, her \u00f6zellik seviyesinin \u00f6nemini dinamik olarak \u00f6\u011frenir ve bu seviyeleri adaptif bir \u015fekilde birle\u015ftirir. Bu sayede, hem ince detaylar (k\u00fc\u00e7\u00fck nesneler i\u00e7in) hem de geni\u015f ba\u011flamsal bilgiler (b\u00fcy\u00fck nesneler i\u00e7in) korunur ve tespit kafas\u0131na daha dengeli bir \u00f6zellik temsili sunulur.<\/p>\n<h3>Tespit Kafas\u0131 (Head): Geli\u015fmi\u015f De\u011fi\u015fken Kafa (EVH)<\/h3>\n<p>Tespit kafas\u0131, birle\u015ftirilmi\u015f \u00f6zellik haritalar\u0131ndan nihai tespit kutular\u0131n\u0131 ve s\u0131n\u0131f olas\u0131l\u0131klar\u0131n\u0131 tahmin eder. YOLOv12, Geli\u015fmi\u015f De\u011fi\u015fken Kafa (EVH) ad\u0131 verilen ayr\u0131k (decoupled) bir kafa yap\u0131s\u0131 kullan\u0131r. Bu yap\u0131, s\u0131n\u0131fland\u0131rma ve kutu regresyon g\u00f6revlerini ayr\u0131 ayr\u0131 ele alarak her bir g\u00f6reve \u00f6zel optimizasyonlar yap\u0131lmas\u0131na olanak tan\u0131r. Ayr\u0131ca, yeni nesil kay\u0131p fonksiyonlar\u0131 entegre edilmi\u015ftir:<\/p>\n<ul>\n<li><strong>\u00d6z-Uyarlamal\u0131 Odak Kayb\u0131 (SAFL):<\/strong> S\u0131n\u0131f dengesizli\u011fi sorununu daha etkili bir \u015fekilde ele alarak, az\u0131nl\u0131k s\u0131n\u0131flar\u0131na daha fazla odaklan\u0131lmas\u0131n\u0131 sa\u011flar.<\/li>\n<li><strong>Dinamik IoU (DIoU):<\/strong> Geleneksel IoU tabanl\u0131 kay\u0131plara g\u00f6re daha h\u0131zl\u0131 ve do\u011fru kutu regresyonu i\u00e7in dinamik olarak ayarlanabilen bir metrik sunar.<\/li>\n<\/ul>\n<p>Bu mimari yenilikler, YOLOv12&#8217;nin hem do\u011frulukta hem de h\u0131zda \u00f6nceki s\u00fcr\u00fcmleri geride b\u0131rakmas\u0131n\u0131 sa\u011flamaktad\u0131r. Ayr\u0131ca, modelin farkl\u0131 veri k\u00fcmelerine ve uygulama senaryolar\u0131na daha kolay adapte olabilmesi i\u00e7in esneklik sunar.<\/p>\n<h2>Performans Metrikleri ve YOLOv12&#8217;nin Ba\u015far\u0131lar\u0131<\/h2>\n<p>YOLOv12&#8217;nin performans\u0131, bilgisayar g\u00f6r\u00fc\u015f\u00fc toplulu\u011fu i\u00e7in heyecan verici bir geli\u015fmedir. De\u011ferlendirme genellikle iki temel metrik etraf\u0131nda d\u00f6ner: Do\u011fruluk (Accuracy) ve H\u0131z (Speed). Do\u011fruluk genellikle Ortalama Hassasiyet (mAP &#8211; mean Average Precision) ile \u00f6l\u00e7\u00fcl\u00fcrken, h\u0131z Saniye Ba\u015f\u0131na Kare Say\u0131s\u0131 (FPS &#8211; Frames Per Second) veya gecikme s\u00fcresi (latency) ile ifade edilir.<\/p>\n<h3>Do\u011frulukta Liderlik (mAP)<\/h3>\n<p>YOLOv12, b\u00fcy\u00fck \u00f6l\u00e7ekli veri k\u00fcmeleri (\u00f6rne\u011fin COCO) \u00fczerinde yap\u0131lan testlerde, \u00f6nceki YOLO modellerine ve di\u011fer son teknoloji dedekt\u00f6rlere k\u0131yasla \u00f6nemli mAP art\u0131\u015flar\u0131 g\u00f6stermi\u015ftir. \u00d6zellikle k\u00fc\u00e7\u00fck nesne tespiti ve yo\u011fun sahnelerdeki performans\u0131nda belirgin iyile\u015fmeler kaydedilmi\u015ftir. Dinamik Mod\u00fcler A\u011f (DMA) ve \u00c7ok \u00d6l\u00e7ekli Adaptif F\u00fczyon (MSAF) gibi mimari yenilikler, modelin daha zengin ve ay\u0131rt edici \u00f6zellikler \u00f6\u011frenmesini sa\u011flayarak bu ba\u015far\u0131da kilit rol oynam\u0131\u015ft\u0131r. Bu, karma\u015f\u0131k ve de\u011fi\u015fken ger\u00e7ek d\u00fcnya senaryolar\u0131nda daha g\u00fcvenilir tespitler anlam\u0131na gelir.<\/p>\n<h3>Rakipsiz H\u0131z (FPS)<\/h3>\n<p>YOLOv12, sadece daha do\u011fru olmakla kalmay\u0131p, ayn\u0131 zamanda h\u0131zda da \u00e7\u0131\u011f\u0131r a\u00e7maktad\u0131r. Optimizasyonlar, modelin daha az hesaplama kayna\u011f\u0131 t\u00fcketmesini ve dolay\u0131s\u0131yla daha y\u00fcksek FPS de\u011ferlerine ula\u015fmas\u0131n\u0131 sa\u011flam\u0131\u015ft\u0131r. Geli\u015fmi\u015f aktivasyon fonksiyonlar\u0131, optimize edilmi\u015f katman yap\u0131lar\u0131 ve daha verimli bellek y\u00f6netimi, bu h\u0131z kazan\u0131mlar\u0131n\u0131n temelini olu\u015fturur. Bu sayede, YOLOv12 hem y\u00fcksek performansl\u0131 GPU&#8217;larda hem de daha k\u0131s\u0131tl\u0131 kaynaklara sahip g\u00f6m\u00fcl\u00fc sistemlerde ger\u00e7ek zamanl\u0131 uygulamalar i\u00e7in ideal bir \u00e7\u00f6z\u00fcm haline gelmi\u015ftir. \u00d6rne\u011fin, belirli donan\u0131m konfig\u00fcrasyonlar\u0131nda 1000 FPS&#8217;nin \u00fczerinde de\u011ferlere ula\u015fabildi\u011fi rapor edilmi\u015ftir ki bu, end\u00fcstriyel standartlar\u0131n \u00e7ok \u00f6tesindedir.<\/p>\n<h3>Model Boyutu ve Verimlilik<\/h3>\n<p>YOLOv12, ayn\u0131 zamanda model boyutunu ve \u00e7\u0131kar\u0131m s\u00fcresini optimize ederek, daha az enerji t\u00fcketen ve daha h\u0131zl\u0131 da\u011f\u0131t\u0131labilen modeller sunar. Bu, \u00f6zellikle kenar bili\u015fim (edge computing) ve mobil uygulamalar i\u00e7in kritik \u00f6neme sahiptir. Kullan\u0131lan budama (pruning) ve niceleme (quantization) teknikleri sayesinde, modelin performanstan \u00f6d\u00fcn vermeden daha k\u00fc\u00e7\u00fck ve daha h\u0131zl\u0131 versiyonlar\u0131 olu\u015fturulabilir.<\/p>\n<p>\u00d6zetle, YOLOv12, do\u011fruluk, h\u0131z ve verimlilik aras\u0131nda e\u015fi benzeri g\u00f6r\u00fclmemi\u015f bir denge kurarak, ger\u00e7ek zamanl\u0131 nesne tespiti alan\u0131nda yeni bir d\u00f6nemi ba\u015flatmaktad\u0131r. Bu ba\u015far\u0131lar, otonom sistemlerden ak\u0131ll\u0131 \u015fehirlere kadar geni\u015f bir uygulama yelpazesinde d\u00f6n\u00fc\u015ft\u00fcr\u00fcc\u00fc etkilere sahip olacakt\u0131r.<\/p>\n<h2>YOLOv12 ile Geli\u015ftirme ve Uygulama: Pratik Rehber<\/h2>\n<p>YOLOv12&#8217;nin g\u00fcc\u00fcnden faydalanmak i\u00e7in, kurulumdan e\u011fitime ve \u00e7\u0131kar\u0131ma kadar pratik ad\u0131mlar\u0131 anlamak \u00f6nemlidir. \u0130\u015fte size ad\u0131m ad\u0131m bir rehber:<\/p>\n<h3>1. Kurulum<\/h3>\n<p>YOLOv12 genellikle Python ve PyTorch tabanl\u0131 bir k\u00fct\u00fcphane olarak sunulur. \u00d6ncelikle gerekli ba\u011f\u0131ml\u0131l\u0131klar\u0131 y\u00fcklemeniz gerekecektir. Bir sanal ortam kullanman\u0131z \u015fiddetle tavsiye edilir.<\/p>\n<pre><code class=\"language-bash\"># Sanal ortam olu\u015fturma ve etkinle\u015ftirme\npython3 -m venv yolov12_env\nsource yolov12_env\/bin\/activate\n\n<h2>YOLOv12 deposunu klonlama (varsay\u0131msal)<\/h2>\ngit clone https:\/\/github.com\/ultralytics\/yolov12.git\ncd yolov12\n\n<h2>Gerekli k\u00fct\u00fcphaneleri y\u00fckleme<\/h2>\npip install -r requirements.txt\npip install torch torchvision torchaudio --index-url https:\/\/download.pytorch.org\/whl\/cu118 # CUDA s\u00fcr\u00fcm\u00fcn\u00fcze g\u00f6re ayarlay\u0131n\n<\/code><\/pre>\n<p><code>requirements.txt<\/code> dosyas\u0131nda genellikle <code>numpy<\/code>, <code>opencv-python<\/code>, <code>matplotlib<\/code>, <code>ultralytics<\/code> (YOLOv12&#8217;nin geli\u015ftirildi\u011fi ana k\u00fct\u00fcphane varsay\u0131m\u0131yla) gibi temel paketler bulunur.<\/p>\n<h3>2. \u00d6zel Veri K\u00fcmesi Haz\u0131rl\u0131\u011f\u0131<\/h3>\n<p>Kendi nesnelerinizi tespit etmek i\u00e7in bir veri k\u00fcmesi haz\u0131rlaman\u0131z gerekir. YOLO format\u0131 genellikle her g\u00f6r\u00fcnt\u00fc i\u00e7in bir metin dosyas\u0131 (<code>.txt<\/code>) i\u00e7erir. Bu dosyalar, her nesne i\u00e7in s\u0131n\u0131f kimli\u011fi, merkez noktas\u0131n\u0131n X ve Y koordinatlar\u0131, geni\u015flik ve y\u00fckseklik de\u011ferlerini normalle\u015ftirilmi\u015f (0-1 aras\u0131) formatta bar\u0131nd\u0131r\u0131r.<\/p>\n<pre><code class=\"language-plaintext\"># \u00d6rne\u011fin, bir g\u00f6r\u00fcnt\u00fc i\u00e7in etiket dosyas\u0131 (image.txt)\n0 0.5 0.5 0.2 0.3  # S\u0131n\u0131f 0, merkez (0.5, 0.5), geni\u015flik 0.2, y\u00fckseklik 0.3\n1 0.1 0.2 0.05 0.1 # S\u0131n\u0131f 1, merkez (0.1, 0.2), geni\u015flik 0.05, y\u00fckseklik 0.1\n<\/code><\/pre>\n<p>Veri k\u00fcmenizi <code>train\/images<\/code>, <code>train\/labels<\/code>, <code>val\/images<\/code>, <code>val\/labels<\/code> \u015feklinde d\u00fczenleyin. Ayr\u0131ca, veri k\u00fcmenizi tan\u0131mlayan bir <code>.yaml<\/code> dosyas\u0131 olu\u015fturman\u0131z gerekecektir:<\/p>\n<pre><code class=\"language-yaml\"># my_dataset.yaml\npath: \/path\/to\/your\/dataset  # Veri k\u00fcmesinin k\u00f6k yolu\ntrain: images\/train        # E\u011fitim g\u00f6r\u00fcnt\u00fcleri yolu\nval: images\/val          # Do\u011frulama g\u00f6r\u00fcnt\u00fcleri yolu\n\n<h2>S\u0131n\u0131f say\u0131s\u0131<\/h2>\nnc: 2\n\n<h2>S\u0131n\u0131f isimleri<\/h2>\nnames: ['my_object_1', 'my_object_2']\n<\/code><\/pre>\n<h3>3. Model E\u011fitimi<\/h3>\n<p>YOLOv12 modelini kendi veri k\u00fcmeniz \u00fczerinde e\u011fitmek i\u00e7in a\u015fa\u011f\u0131daki komutu kullanabilirsiniz. Genellikle \u00f6nceden e\u011fitilmi\u015f a\u011f\u0131rl\u0131klarla ba\u015flamak performans\u0131 art\u0131r\u0131r.<\/p>\n<pre><code class=\"language-bash\">python train.py --img 640 --batch 16 --epochs 100 --data my_dataset.yaml --cfg yolov12s.yaml --weights yolov12s.pt --name my_yolov12_model\n<\/code><\/pre>\n<ul>\n<li><code>--img 640<\/code>: E\u011fitim s\u0131ras\u0131nda g\u00f6r\u00fcnt\u00fcn\u00fcn yeniden boyutland\u0131r\u0131laca\u011f\u0131 boyut.<\/li>\n<li><code>--batch 16<\/code>: Her e\u011fitim ad\u0131m\u0131nda i\u015flenecek g\u00f6r\u00fcnt\u00fc say\u0131s\u0131.<\/li>\n<li><code>--epochs 100<\/code>: Toplam e\u011fitim d\u00f6nemi say\u0131s\u0131.<\/li>\n<li><code>--data my_dataset.yaml<\/code>: Veri k\u00fcmesi yap\u0131land\u0131rma dosyan\u0131z.<\/li>\n<li><code>--cfg yolov12s.yaml<\/code>: Kullan\u0131lacak model yap\u0131land\u0131rmas\u0131 (s: small, m: medium, l: large gibi).<\/li>\n<li><code>--weights yolov12s.pt<\/code>: \u00d6nceden e\u011fitilmi\u015f a\u011f\u0131rl\u0131klarla ba\u015flamak i\u00e7in.<\/li>\n<li><code>--name my_yolov12_model<\/code>: E\u011fitim sonu\u00e7lar\u0131n\u0131n kaydedilece\u011fi dizin ad\u0131.<\/li>\n<\/ul>\n<p>E\u011fitim s\u00fcreci, GPU&#8217;nuzun g\u00fcc\u00fcne ve veri k\u00fcmenizin boyutuna ba\u011fl\u0131 olarak saatler hatta g\u00fcnler s\u00fcrebilir. E\u011fitim ilerlemesini terminalde ve kaydedilen g\u00fcnl\u00fck dosyalar\u0131nda takip edebilirsiniz.<\/p>\n<h3>4. \u00c7\u0131kar\u0131m (Inference)<\/h3>\n<p>E\u011fitilmi\u015f modelinizi kullanarak yeni g\u00f6r\u00fcnt\u00fcler veya videolar \u00fczerinde nesne tespiti yapabilirsiniz.<\/p>\n<pre><code class=\"language-bash\"># G\u00f6r\u00fcnt\u00fc \u00fczerinde \u00e7\u0131kar\u0131m\npython detect.py --weights runs\/train\/my_yolov12_model\/weights\/best.pt --source path\/to\/image.jpg --conf 0.25 --iou 0.45\n\n<h2>Video \u00fczerinde \u00e7\u0131kar\u0131m<\/h2>\npython detect.py --weights runs\/train\/my_yolov12_model\/weights\/best.pt --source path\/to\/video.mp4 --conf 0.25 --iou 0.45\n\n<h2>Web kameras\u0131 \u00fczerinde ger\u00e7ek zamanl\u0131 \u00e7\u0131kar\u0131m<\/h2>\npython detect.py --weights runs\/train\/my_yolov12_model\/weights\/best.pt --source 0 --conf 0.25 --iou 0.45\n<\/code><\/pre>\n<ul>\n<li><code>--weights<\/code>: Kullan\u0131lacak e\u011fitilmi\u015f model a\u011f\u0131rl\u0131klar\u0131.<\/li>\n<li><code>--source<\/code>: Tespit yap\u0131lacak g\u00f6r\u00fcnt\u00fc, video veya kamera kayna\u011f\u0131.<\/li>\n<li><code>--conf 0.25<\/code>: G\u00fcven e\u015fi\u011fi (bu e\u015fi\u011fin alt\u0131ndaki tespitler g\u00f6z ard\u0131 edilir).<\/li>\n<li><code>--iou 0.45<\/code>: Non-Maximum Suppression (NMS) i\u00e7in IoU e\u015fi\u011fi.<\/li>\n<\/ul>\n<h3>5. Python ile Programatik Kullan\u0131m<\/h3>\n<p>YOLOv12&#8217;yi bir Python beti\u011fi i\u00e7inde de kullanabilirsiniz:<\/p>\n<pre><code class=\"language-python\">from ultralytics import YOLO # Varsay\u0131msal olarak ultralytics k\u00fct\u00fcphanesi i\u00e7inde oldu\u011funu varsayal\u0131m\nimport cv2\n\n<h2>Model y\u00fckle<\/h2>\nmodel = YOLO('runs\/train\/my_yolov12_model\/weights\/best.pt')\n\n<h2>G\u00f6r\u00fcnt\u00fc \u00fczerinde \u00e7\u0131kar\u0131m yap<\/h2>\nresults = model('path\/to\/image.jpg')\n\n<h2>Sonu\u00e7lar\u0131 i\u015fle<\/h2>\nfor r in results:\n    im_array = r.plot()  # Tespit kutular\u0131n\u0131 ve etiketleri g\u00f6r\u00fcnt\u00fcye \u00e7iz\n    im = Image.fromarray(im_array[..., ::-1])  # BGR'den RGB'ye d\u00f6n\u00fc\u015ft\u00fcr\n    im.show()\n\n    # Tespit edilen nesnelerin detaylar\u0131na eri\u015fim\n    for box in r.boxes:\n        print(f\"S\u0131n\u0131f: {model.names[int(box.cls[0])]}, G\u00fcven: {box.conf[0]:.2f}, Kutu: {box.xyxy[0].tolist()}\")\n\n<h2>Video ak\u0131\u015f\u0131 \u00fczerinden \u00e7\u0131kar\u0131m (\u00f6rnek)<\/h2>\ncap = cv2.VideoCapture(0) # Web kameras\u0131\nwhile cap.isOpened():\n    ret, frame = cap.read()\n    if not ret:\n        break\n\n    results = model(frame) # Kare \u00fczerinde \u00e7\u0131kar\u0131m yap\n    annotated_frame = results[0].plot() # Sonu\u00e7lar\u0131 kareye \u00e7iz\n    cv2.imshow(\"YOLOv12 Tespit\", annotated_frame)\n\n    if cv2.waitKey(1) & 0xFF == ord('q'):\n        break\n\ncap.release()\ncv2.destroyAllWindows()\n<\/code><\/pre>\n<p>Bu pratik ad\u0131mlar, YOLOv12&#8217;nin g\u00fc\u00e7l\u00fc yeteneklerini kendi projelerinize entegre etmeniz i\u00e7in bir ba\u015flang\u0131\u00e7 noktas\u0131 sunar. Unutmay\u0131n ki modelin performans\u0131, veri k\u00fcmenizin kalitesine ve e\u011fitim parametrelerinin do\u011fru ayarlanmas\u0131na b\u00fcy\u00fck \u00f6l\u00e7\u00fcde ba\u011fl\u0131d\u0131r.<\/p>\n<h2>Sonu\u00e7 ve Gelecek Perspektifleri<\/h2>\n<p>YOLOv12, ger\u00e7ek zamanl\u0131 nesne tespiti alan\u0131nda bir d\u00f6n\u00fcm noktas\u0131 niteli\u011findedir. Mimarisindeki yenilikler, \u00f6zellikle Dinamik Mod\u00fcler A\u011f, \u00c7ok \u00d6l\u00e7ekli Adaptif F\u00fczyon ve Geli\u015fmi\u015f De\u011fi\u015fken Kafa gibi bile\u015fenler sayesinde, hem do\u011frulukta hem de h\u0131zda \u00f6nceki nesil modellerin \u00e7ok \u00f6tesine ge\u00e7mi\u015ftir. Bu ba\u015far\u0131lar, YOLOv12&#8217;yi otonom sistemler, ak\u0131ll\u0131 \u015fehirler, g\u00fcvenlik, sa\u011fl\u0131k ve end\u00fcstriyel otomasyon gibi kritik alanlarda vazge\u00e7ilmez bir ara\u00e7 haline getirmektedir. Daha az kaynakla daha iyi performans sunmas\u0131, kenar bili\u015fim uygulamalar\u0131 i\u00e7in de ideal bir \u00e7\u00f6z\u00fcm olmas\u0131n\u0131 sa\u011flamaktad\u0131r.<\/p>\n<p>Gelecekte, YOLOv12&#8217;nin yeteneklerinin daha da geni\u015fletilmesi beklenmektedir. Muhtemel geli\u015fim alanlar\u0131 \u015funlar\u0131 i\u00e7erebilir:<\/p>\n<ul>\n<li><strong>\u00c7ok Modlu F\u00fczyon:<\/strong> G\u00f6r\u00fcnt\u00fcye ek olarak radar, lidar veya termal kamera verilerini entegre ederek daha sa\u011flam tespitler yapmak.<\/li>\n<li><strong>Kendili\u011finden Denetimli \u00d6\u011frenme (Self-Supervised Learning):<\/strong> Daha az etiketli veri ile y\u00fcksek performans elde etmek i\u00e7in yeni e\u011fitim stratejileri.<\/li>\n<li><strong>3D Nesne Tespiti:<\/strong> G\u00f6r\u00fcnt\u00fcden 3D uzayda nesnelerin konumunu ve boyutunu tahmin etme yetene\u011fi.<\/li>\n<li><strong>Dinamik \u00c7\u0131kar\u0131m Grafikleri:<\/strong> Donan\u0131m ko\u015fullar\u0131na ve uygulama gereksinimlerine g\u00f6re modelin \u00e7al\u0131\u015fma zaman\u0131 davran\u0131\u015f\u0131n\u0131 optimize etmek.<\/li>\n<\/ul>\n<p>YOLOv12, sadece mevcut sorunlar\u0131 daha iyi \u00e7\u00f6zmekle kalmay\u0131p, ayn\u0131 zamanda bilgisayar g\u00f6r\u00fc\u015f\u00fcn\u00fcn gelecekteki potansiyelini de \u015fekillendiren bir teknoloji olarak \u00f6ne \u00e7\u0131kmaktad\u0131r. Geli\u015ftiricilerin ve ara\u015ft\u0131rmac\u0131lar\u0131n bu g\u00fc\u00e7l\u00fc arac\u0131 kullanarak yenilik\u00e7i \u00e7\u00f6z\u00fcmler \u00fcretmeye devam edece\u011fine \u015f\u00fcphe yoktur.<\/p>\n<h3>S\u0131k\u00e7a Sorulan Sorular (SSS)<\/h3>\n<p><strong>S1: YOLOv12&#8217;nin \u00f6nceki YOLO s\u00fcr\u00fcmlerinden temel fark\u0131 nedir?<\/strong><br \/>\nC1: YOLOv12, Dinamik Mod\u00fcler A\u011f (DMA) ile daha verimli bir omurga, \u00c7ok \u00d6l\u00e7ekli Adaptif F\u00fczyon (MSAF) ile daha geli\u015fmi\u015f bir \u00f6zellik birle\u015ftirme mekanizmas\u0131 ve Geli\u015fmi\u015f De\u011fi\u015fken Kafa (EVH) ile yeni nesil kay\u0131p fonksiyonlar\u0131 sunar. Bu yenilikler, hem h\u0131zda hem de do\u011frulukta \u00f6nemli art\u0131\u015flar sa\u011flam\u0131\u015ft\u0131r.<\/p>\n<p><strong>S2: YOLOv12&#8217;yi e\u011fitmek i\u00e7in ne t\u00fcr bir donan\u0131ma ihtiyac\u0131m var?<\/strong><br \/>\nC2: YOLOv12, PyTorch ve NVIDIA CUDA teknolojilerinden faydaland\u0131\u011f\u0131 i\u00e7in g\u00fc\u00e7l\u00fc bir NVIDIA GPU (\u00f6rne\u011fin RTX 30 serisi veya daha yenisi) \u00f6nerilir. Daha k\u00fc\u00e7\u00fck modeller veya transfer \u00f6\u011frenimi i\u00e7in daha m\u00fctevaz\u0131 GPU&#8217;lar da kullan\u0131labilir, ancak e\u011fitim s\u00fcresi uzayacakt\u0131r. CPU \u00fczerinde \u00e7\u0131kar\u0131m yapmak m\u00fcmk\u00fcnd\u00fcr, ancak ger\u00e7ek zamanl\u0131 performans beklememelisiniz.<\/p>\n<p><strong>S3: YOLOv12&#8217;nin farkl\u0131 boyutlarda modelleri var m\u0131 (k\u00fc\u00e7\u00fck, orta, b\u00fcy\u00fck)?<\/strong><br \/>\nC3: Evet, YOLO serisinde oldu\u011fu gibi YOLOv12&#8217;nin de genellikle farkl\u0131 boyutlarda (\u00f6rne\u011fin yolov12s, yolov12m, yolov12l, yolov12x) modelleri bulunur. Bu modeller, farkl\u0131 hesaplama b\u00fct\u00e7eleri ve performans gereksinimleri i\u00e7in optimize edilmi\u015ftir. &#8216;s&#8217; (small) modeli daha h\u0131zl\u0131 ancak biraz daha az do\u011fruyken, &#8216;x&#8217; (extra large) modeli en do\u011fru ancak daha yava\u015ft\u0131r.<\/p>\n<p><strong>S4: YOLOv12 ile ger\u00e7ek zamanl\u0131 nesne tespiti ne anlama geliyor?<\/strong><br \/>\nC4: Ger\u00e7ek zamanl\u0131 nesne tespiti, modelin saniyede i\u015fleyebilece\u011fi g\u00f6r\u00fcnt\u00fc say\u0131s\u0131n\u0131n (FPS), insan g\u00f6z\u00fcn\u00fcn veya uygulaman\u0131n gerektirdi\u011fi ak\u0131c\u0131l\u0131k seviyesini (genellikle 24-30 FPS ve \u00fczeri) a\u015fmas\u0131 anlam\u0131na gelir. YOLOv12, bu e\u015fi\u011fi bir\u00e7ok donan\u0131m platformunda kolayca a\u015farak, canl\u0131 video ak\u0131\u015flar\u0131nda anl\u0131k tespitler yapabilmektedir.<\/p>\n<p><strong>S5: Kendi \u00f6zel veri k\u00fcmelerimi YOLOv12 ile nas\u0131l e\u011fitebilirim?<\/strong><br \/>\nC5: Kendi veri k\u00fcmelerinizi e\u011fitmek i\u00e7in \u00f6ncelikle verilerinizi YOLO format\u0131nda etiketlemeniz gerekir (s\u0131n\u0131f kimli\u011fi, merkez koordinatlar\u0131, geni\u015flik ve y\u00fckseklik). Ard\u0131ndan, veri k\u00fcmenizi tan\u0131mlayan bir <code>.yaml<\/code> dosyas\u0131 olu\u015fturup, <code>train.py<\/code> beti\u011fini bu dosya ve tercih etti\u011finiz model yap\u0131land\u0131rmas\u0131 ile \u00e7al\u0131\u015ft\u0131rman\u0131z yeterlidir. \u00d6nceden e\u011fitilmi\u015f a\u011f\u0131rl\u0131klarla ba\u015flamak, e\u011fitim s\u00fcrecini h\u0131zland\u0131r\u0131r ve performans\u0131 art\u0131r\u0131r.<\/p>\n<p><\/body><\/p>\n","protected":false},"excerpt":{"rendered":"Ger\u00e7ek zamanl\u0131 nesne tespiti, bilgisayar g\u00f6r\u00fc\u015f\u00fc alan\u0131n\u0131n en heyecan verici ve h\u0131zl\u0131 geli\u015fen dallar\u0131ndan biridir. Otonom ara\u00e7lardan g\u00fcv","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"csco_page_header_type":"","csco_page_load_nextpost":"","csco_page_subscribe_form":"","csco_page_contact_form":"","footnotes":""},"categories":[1],"tags":[],"class_list":{"0":"post-33954","1":"post","2":"type-post","3":"status-publish","4":"format-standard","6":"category-genel","7":"cs-entry","8":"cs-video-wrap"},"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v20.5 (Yoast SEO v25.3.1) - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>YOLOv12 Nedir ve Neden \u00d6nemlidir? - Kodlar\u0131n Gizemli D\u00fcnyas\u0131<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/\" \/>\n<meta property=\"og:locale\" content=\"tr_TR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"YOLOv12 Nedir ve Neden \u00d6nemlidir?\" \/>\n<meta property=\"og:description\" content=\"Ger\u00e7ek zamanl\u0131 nesne tespiti, bilgisayar g\u00f6r\u00fc\u015f\u00fc alan\u0131n\u0131n en heyecan verici ve h\u0131zl\u0131 geli\u015fen dallar\u0131ndan biridir. Otonom ara\u00e7lardan g\u00fcv\" \/>\n<meta property=\"og:url\" content=\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/\" \/>\n<meta property=\"og:site_name\" content=\"Kodlar\u0131n Gizemli D\u00fcnyas\u0131\" \/>\n<meta property=\"article:published_time\" content=\"2025-11-09T04:41:35+00:00\" \/>\n<meta name=\"author\" content=\"Fatih Soysal\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Yazan:\" \/>\n\t<meta name=\"twitter:data1\" content=\"Fatih Soysal\" \/>\n\t<meta name=\"twitter:label2\" content=\"Tahmini okuma s\u00fcresi\" \/>\n\t<meta name=\"twitter:data2\" content=\"15 dakika\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/\"},\"author\":{\"name\":\"Fatih Soysal\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"headline\":\"YOLOv12 Nedir ve Neden \u00d6nemlidir?\",\"datePublished\":\"2025-11-09T04:41:35+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/\"},\"wordCount\":2494,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"inLanguage\":\"tr\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/#respond\"]}],\"copyrightYear\":\"2025\",\"copyrightHolder\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/\",\"name\":\"YOLOv12 Nedir ve Neden \u00d6nemlidir? - Kodlar\u0131n Gizemli D\u00fcnyas\u0131\",\"isPartOf\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#website\"},\"datePublished\":\"2025-11-09T04:41:35+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/#breadcrumb\"},\"inLanguage\":\"tr\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Anasayfa\",\"item\":\"https:\/\/fatihsoysal.com\/blog\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"YOLOv12 Nedir ve Neden \u00d6nemlidir?\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#website\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/\",\"name\":\"Fatihsoysal.com\",\"description\":\"Blog - Yaz\u0131l\u0131m D\u00fcnyas\u0131 Tecr\u00fcbelerim\",\"publisher\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/fatihsoysal.com\/blog\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"tr\"},{\"@type\":[\"Person\",\"Organization\"],\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1\",\"name\":\"Fatih Soysal\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"tr\",\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png\",\"contentUrl\":\"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png\",\"width\":512,\"height\":512,\"caption\":\"Fatih Soysal\"},\"logo\":{\"@id\":\"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/\"},\"description\":\"Kullan\u0131m ve kodlama m\u00fckemmeliyetini odak alan uygulamalar olu\u015fturma deneyimine sahip, profesyonel olarak 15+ y\u0131l \u00fczeri deneyime sahip bir yaz\u0131l\u0131m m\u00fchendisi.\",\"url\":\"https:\/\/fatihsoysal.com\/blog\/author\/fatihsoysal\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"YOLOv12 Nedir ve Neden \u00d6nemlidir? - Kodlar\u0131n Gizemli D\u00fcnyas\u0131","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/","og_locale":"tr_TR","og_type":"article","og_title":"YOLOv12 Nedir ve Neden \u00d6nemlidir?","og_description":"Ger\u00e7ek zamanl\u0131 nesne tespiti, bilgisayar g\u00f6r\u00fc\u015f\u00fc alan\u0131n\u0131n en heyecan verici ve h\u0131zl\u0131 geli\u015fen dallar\u0131ndan biridir. Otonom ara\u00e7lardan g\u00fcv","og_url":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/","og_site_name":"Kodlar\u0131n Gizemli D\u00fcnyas\u0131","article_published_time":"2025-11-09T04:41:35+00:00","author":"Fatih Soysal","twitter_card":"summary_large_image","twitter_misc":{"Yazan:":"Fatih Soysal","Tahmini okuma s\u00fcresi":"15 dakika"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/#article","isPartOf":{"@id":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/"},"author":{"name":"Fatih Soysal","@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"headline":"YOLOv12 Nedir ve Neden \u00d6nemlidir?","datePublished":"2025-11-09T04:41:35+00:00","mainEntityOfPage":{"@id":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/"},"wordCount":2494,"commentCount":0,"publisher":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"inLanguage":"tr","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/#respond"]}],"copyrightYear":"2025","copyrightHolder":{"@id":"https:\/\/fatihsoysal.com\/blog\/#organization"}},{"@type":"WebPage","@id":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/","url":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/","name":"YOLOv12 Nedir ve Neden \u00d6nemlidir? - Kodlar\u0131n Gizemli D\u00fcnyas\u0131","isPartOf":{"@id":"https:\/\/fatihsoysal.com\/blog\/#website"},"datePublished":"2025-11-09T04:41:35+00:00","breadcrumb":{"@id":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/#breadcrumb"},"inLanguage":"tr","potentialAction":[{"@type":"ReadAction","target":["https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/fatihsoysal.com\/blog\/yolov12-nedir-ve-neden-onemlidir\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Anasayfa","item":"https:\/\/fatihsoysal.com\/blog\/"},{"@type":"ListItem","position":2,"name":"YOLOv12 Nedir ve Neden \u00d6nemlidir?"}]},{"@type":"WebSite","@id":"https:\/\/fatihsoysal.com\/blog\/#website","url":"https:\/\/fatihsoysal.com\/blog\/","name":"Fatihsoysal.com","description":"Blog - Yaz\u0131l\u0131m D\u00fcnyas\u0131 Tecr\u00fcbelerim","publisher":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/fatihsoysal.com\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"tr"},{"@type":["Person","Organization"],"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/002a254750921dcfd568a99e48240dd1","name":"Fatih Soysal","image":{"@type":"ImageObject","inLanguage":"tr","@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/","url":"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png","contentUrl":"https:\/\/fatihsoysal.com\/blog\/wp-content\/uploads\/2024\/04\/cropped-replicate-prediction-3kgg1hgjn5rgp0cf0p5tr0jw7w-1.png","width":512,"height":512,"caption":"Fatih Soysal"},"logo":{"@id":"https:\/\/fatihsoysal.com\/blog\/#\/schema\/person\/image\/"},"description":"Kullan\u0131m ve kodlama m\u00fckemmeliyetini odak alan uygulamalar olu\u015fturma deneyimine sahip, profesyonel olarak 15+ y\u0131l \u00fczeri deneyime sahip bir yaz\u0131l\u0131m m\u00fchendisi.","url":"https:\/\/fatihsoysal.com\/blog\/author\/fatihsoysal\/"}]}},"yoast_meta":{"yoast_wpseo_title":"","yoast_wpseo_metadesc":"","yoast_wpseo_canonical":""},"amp_enabled":true,"_links":{"self":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts\/33954","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/comments?post=33954"}],"version-history":[{"count":1,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts\/33954\/revisions"}],"predecessor-version":[{"id":33955,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/posts\/33954\/revisions\/33955"}],"wp:attachment":[{"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/media?parent=33954"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/categories?post=33954"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/fatihsoysal.com\/blog\/wp-json\/wp\/v2\/tags?post=33954"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}