The classification models mainly based on CBAM-keras, the others are models that i explore and try:
- Frame include tensorflow and pytorch
- Specific details you can view README.md or TRAINING.md under each floder
- the structure of CBAM : Convolutional Block Attention Module"
- The main project file I use =
./ImageClassClassification/CBAM-tensorflow-slim
- Inception V4 + CBAM / + SE
- Inception-ResNet-v2 + CBAM / + SE
- ResNet V1 50 + CBAM / + SE
- ResNet V1 101 + CBAM / + SE
- ResNet V1 152 + CBAM / + SE
- ResNet V1 200 + CBAM / + SE
- ResNet V2 50 + CBAM / + SE
- ResNet V2 101 + CBAM / + SE
- ResNet V2 152 + CBAM / + SE
- ResNet V2 200 + CBAM / + SE
- Python 3.x
- TensorFlow 1.x
- TF-slim
- torch 1.x
- Keras (IMDB dataset)
- tqdm
- scikit-image
- numpy
- torch>=0.4.0
- torchvision
- pillow
- matplotlib
- wing
The application scenario of the model is vehicle detection.There are SSD and yolov3
In the project, the model i used is MyYOLO. Other floders are versions on keras and pytorch.
- Result: When the confidence is 0.8, the accuracy rate is above 0.95.
- SSD is an unified framework for object detection with a single network. It has been originally introduced in this research article.
- YOLOv3 [Original Implementation]
- The main project file I use =
./TargetDetection/MyYOLO
Semantic segmentation in cable quality inspection and autonomous driving applications Irregular boundary lines by semantic segmentation to achieve quality inspection the main network in the project is Bisenet+resnet50
- network structure, aritcle of Bisenet
- The main project file I use =
./SemantemeDivision/Segmentation - Supported models:
- Fontends
- Inceptions_v4
- Mobilenet_v2
- Resnet_v1
- Resnet_v2
- Se_resnext
- Builders:
- Bisenet
- Deeplab_v3
- Refinenet
- Fontends