Starflower8
commented
9 months ago Hello @b03505036,
Thank you for raising this issue and informing us about the updates needed for citation format in the project. Adapting a universally accepted citation style such as APA, MLA, or Chicago indeed helps in preserving consistency and enhancing the project's readability.
However, if I may suggest, adopting the APA (American Psychological Association) referencing style could potentially align more closely with the nature of our project, which primarily resides within the realm of technical and scientific writing. The APA format is commonly used in these disciplines and should cater to the needs of most of our contributors and readers.
Here's how the same citation you provided would look in APA format:
Ou, Y., Yang, R., Ma, L., Liu, Y., Yan, J., Xu, S., Wang, C., & Li, X. (2022). UniInst: Unique representation for end-to-end instance segmentation. Neurocomputing, 514, 551-562. https://doi.org/10.1016/j.neucom.2022.09.112
Do note that in APA style, only the first word of a title or subtitle, proper nouns, or initials are capitalized. The volume number is italicized but not the issue number. The page number range follows the volume and issue details.
However, I also value the thoughts and suggestions of other contributors, and encourage everyone to voice their suggestions or preferences regarding this matter.
Best Regards,
jerry0519
I have received requests to update the citation format in our project to a more standard style, such as APA, MLA, or Chicago. This change is to enhance readability and consistency. I will also update our contributor guidelines to reflect this requirement. Any suggestions or preferences for citation styles are welcome.
@article{OU2022551, title = {UniInst: Unique representation for end-to-end instance segmentation}, journal = {Neurocomputing}, volume = {514}, pages = {551-562}, year = {2022}, issn = {0925-2312}, doi = {https://doi.org/10.1016/j.neucom.2022.09.112}, url = {https://www.sciencedirect.com/science/article/pii/S0925231222012048}, author = {Yimin Ou and Rui Yang and Lufan Ma and Yong Liu and Jiangpeng Yan and Shang Xu and Chengjie Wang and Xiu Li}, keywords = {Instance segmentation, End-to-end instance segmentation, Fully convolutional networks}, abstract = {Existing instance segmentation methods have achieved impressive performance but still suffer from a common dilemma: redundant representations (e.g., multiple boxes, grids, and anchor points) are inferred for one instance, which leads to multiple duplicated predictions. Thus, mainstream methods usually rely on a hand-designed non-maximum suppression (NMS) post-processing step to select the optimal prediction result, consequently hindering end-to-end training. To address this issue, we propose a box-free and NMS-free end-to-end instance segmentation framework, dubbed UniInst, which yields only one unique representation for each instance. Specifically, we design an instance-aware one-to-one assignment scheme, named Only Yield One Representation (OYOR). It dynamically assigns one unique representation to each instance according to the matching quality between predictions and ground truths. Then, a novel prediction re-ranking strategy is elegantly integrated into the framework to address the misalignment between the classification score and mask quality, enabling the learned representation to be more discriminative. With these techniques, our UniInst, the first FCN-based box-free and NMS-free end-to-end instance segmentation framework, achieves competitive performance, e.g., 39.0 mask AP using ResNet-50-FPN and 40.2 mask AP using ResNet-101-FPN on COCO test-dev. Moreover, the proposed instance-aware method is robust to occlusion scenes because of non-dependent on box and NMS. It outperforms common baselines by remarkable mask AP on the heavily-occluded OCHuman benchmark. Code is available at https://github.com/b03505036/UniInst.} }