I need a CNN based image segmentation model including the pre-processing code, the training code, test code and inference code. The image segmentation model shall be trained on a dataset that I can provide.
The basic principle of what the model shall do is depicted in '[login to view URL]'
One example of a image pair from the taining set (input image + output segmentation) is given in '[login to view URL]'
The training code shall also include data augmentation with following steps:
- random image rotation
- random image cropping
- random color shifting
- adding noise in different levels
- adding blur to the image in different levels
- adding a random background image (from a dataset like ImageNet dataset)
The model architecture shall be based on an encoder - decoder network. The decoder backbone shall be a MobileNetV2 (or or possibly newer if it benefits the performance) and the decoder backbone shall be a DeepLabV3. The intended model architecture is depicted in '[login to view URL]'
16 pekerja bebas membida secara purata €137 untuk pekerjaan ini
your task was interesting . I had a experience in computer vision and also very well know about data augmentation techniques. i am new to the freelancer and if your ok with my bid kindly ping me.