Aesthetic Quality Assessment on Photo.net
65.2AccuracyMTCNN #1_FT
Evaluation results
| Method | Links | |
|---|---|---|
| MTCNN #1_FTdelta (δ)=0, Training mode=fine-tuning, Pre-trained on=AVA dataset, Backbone=Multi-task CNN2016.04 | 65.2 | |
| STCNN_FTdelta (δ)=0, Training mode=fine-tuning, Pre-trained on=AVA dataset, Backbone=Single-task CNN2016.04 | 62.1 | |
| STCNNdelta (δ)=0, Training mode=from scratch, Backbone=Single-task CNN2016.04 | 61 | |
| FV_SIFT_SVMdelta (δ)=0, Feature descriptor=FV + SIFT, Classifier=SVM2016.04 | 60.8 | |
| GIST_SVMdelta (δ)=0, Feature descriptor=GIST, Classifier=SVM2016.04 | 59.9 |