https://arxiv.org/pdf/1706.08249.pdf
Few-Example Object Detection with Model Communication，Xuanyi Dong, Liang Zheng, Fan Ma, Yi Yang, Deyu Meng
亮點
- 本文僅僅通過每個類別3-4個bounding box標注即可實現物體檢測，並與其它使用大量training examples的方法性能可比
- 主要方法是：multi-modal learning (多模型同時訓練) ＋ self-paced learning (curriculum learning)
相關工作
這里介紹幾個比較容易混淆的概念，以及與他們相關的方法
- 弱監督物體檢測：數據集的標簽是不可靠的，如（x，y），y對於x的標記是不可靠的。這里的不可靠可以是標記不正確，多種標記，標記不充分，局部標記等。
- 標簽是圖像級別的類別標簽[7][8][9][10][11][18][30][31][32][33][34]
- 半監督物體檢測：半監督學習使用大量的未標記數據，以及同時使用標記數據，來進行模式識別工作。
- 一些訓練樣本只有類別標簽，另外一些樣本有詳細的物體框和類別標注[4][5][6]
- 需要大量標注 (e.g., 50% of the full annotations)
- 每個類別只有幾個物體框標注（Few-Example Object Detection with Model Communication)[12][35]
- 和few-shot learning 的區別：是否使用未標注數據學習
- 通過視頻挖掘位置標注，此類方法主要針對會移動的物體[2][3][29][1]
- Webly supervised learning for object detection: reduce the annotation cost by leveraging web data
方法
Basic detector: Faster RCNN & RFCN
Object proposal method: selective search & edge boxes
Annotations: when we randomly annotate approximately four images for each class, an image may contain several objects, and we annotate all the object bounding boxes.
參數更新：
更新vj：對上述損失函數進行求導，可以得到vj的解
對同一張圖像i同一個模型j，如果有多個樣本使得vj＝1，則只選擇使Lc最小的那個樣本置為1，其他置為0。gamma促使模型之間共享信息，因為vj為1時，閾值變大，圖像更容易被選擇到。
更新wj：與其它文章方法相同
更新yuj：為更新yuj我們需要從一組bounding box找到滿足以下條件的解，
很難直接找到最優化的解。文中采用的方案是：將所有模型預測出的結果輸入nms，並通過閾值只保留分數高的結果，余下的組成yuj。
去除難例：we employ a modified NMS (intersection/max(area1,area2)) to filter out the nested boxes, which usually occurs when there are multiple overlapping objects. If there are too many boxes (≥ 4) for one specific class or too many classes (≥ 4) in the image, this image will be removed. Images in which no reliable pseudo objects are found are filtered out.
實驗
Compared with the-state-of-the-art (4.2 images per class is annotated)
- VOC 2007: -1.1mAP, correct localization +0.9% compared with [21]
- VOC 2012: -2.5mAP compared with [21], correct localization +9.8%
- ILSVRC 2013: -2.4mAP compared with [21]
- COCO 2014: +1.3 mAP compared with [22]
[20] V. Kantorov, M. Oquab, M. Cho, and I. Laptev, “Contextlocnet: Context-aware deep network models for weakly supervised localization,” in European Conference on Computer Vision, 2016.
[21] A. Diba, V. Sharma, A. Pazandeh, H. Pirsiavash, and L. Van Gool, “Weakly supervised cascaded convolutional networks,” 2017
[22] Y. Zhu, Y. Zhou, Q. Ye, Q. Qiu, and J. Jiao, “Soft proposal networks for weakly supervised object localization,” in International Conference on Computer Vision, 2017.
Ablation study
- VOC 2007: +4.1 mAP compared with model ensemble
- k number of labeled images per class; w/ image labels: image-level supervision incorporated
不足
雖然localization有一定准確率，但是難例圖片漏檢比較多（也就是說few example classification效果不好）。