Instance Segmentation of Vehicle Images


Need help with training vehicle image model similar to the existing Mask RCNN pre-trained model (mask_rcnn_coco.h5).

I tried running this model , mask_rcnn_coco.h5, against several vehicle images in order to produce vehicles minus the background. The resulting images were rough outlines of the vehicle. Need help getting cleaner results. Refer to image above to see an example of results.

Willing to hire a professional to achieve good working model better than “mask_rcnn_coco.h5”.

Thank you!
Jose Correa