RetinaMask is a one-stage object detection method that improves upon RetinaNet by adding the task of instance mask prediction during training, as well as an adaptive loss that improves robustness to parameter choice during training, and including more difficult examples in training.
Source: RetinaMask: Learning to predict masks improves state-of-the-art single-shot detection for freePaper | Code | Results | Date | Stars |
---|
Component | Type |
|
---|---|---|
Convolution
|
Convolutions | |
FPN
|
Feature Extractors | |
ReLU
|
Activation Functions | |
RetinaNet
|
Object Detection Models | |
RoIAlign
|
RoI Feature Extractors | |
Self-Adjusting Smooth L1 Loss
|
Loss Functions |