Apply if you have experience with Mask SSD Mobilenet and RCNN.
1. Combine SSD MobileNet model with FCN/SegNet model to make an instance segmentation model like Mask SSD to handle occlusions while object detection.
2. Improve Mask SSD's accuracy and bring it as close as possible to Mask RCNN's accuracy
3. Improve SSD MobileNet's (object detection) accuracy and bring it as close as possible to the Faster RCNN's accuracy
Mask SSD model with code and documentationCode for improved Mask
SSD accuracy with compared accuracy between Mask SSD and Mask RCNN along with documentationCode for improved SSD
MobileNet accuracy with compared accuracy between SSD MobileNet and Faster RCNN along with documentation
Dataset - You can see this link ([login to view URL]) for downloading COCO dataset for panoptic segmentation(2017 Val images [5K/1GB] and 2017 Panoptic Train/Val annotations [821MB]). We csn use partial data if it too big.
5 freelancers are bidding on average ₹10556 for this job
I am a computer vision expert and i already worked on object detection and image segmentation projects and also have experience in ssd mobilenet. So please ping me for more details am perfect fit for your requirements.