物件搜尋在電腦視覺的領域中是一個極難的問題，這個問題的目標主要是在於從給予的圖片中，找出共有的物體。若給予一些圖片，我們所要做的是找出適當的框去標記共同物體。對於共同的物體可能有型態上變化，例如大小、姿勢、…
、外觀等。這些變化會使物件搜尋這個問題變得更佳困難。卷積神經網絡有助於人們方便解決電腦視覺問題。在此文中，我們將會展是一種非監督是方法，來幫助我們學習有效的特徵，並運用在物件搜尋這的問題。

The task of object discovery is to gure out common categories of objects in multiple images without prior knowledge of object categories. It is considered as very challenging computer vision problem. Given a set of images, we aim to identify and localize the common objects. The common objects may vary in scales, poses, appearances, and with occlusions, and these variations make the task of object discovery more diffcult. Conventional solutions tackle the problem with the aid of human intervention. In this work, we present an unsupervised method to learn effective features for object
discovery.

1 Introduction 9
2 Related Work 11
3 The Proposed Method 13
3.1 Unsupervised feature learning 13
3.1.1 Training data preparing 14
3.1.2 Unsupervised feature learning model 17
3.2 Object discovery 18
3.2.1 Supported relation 19
3.2.2 Refining the matches of proposals 20
3.2.3 Region selection 22
4 Experiments 24
4.1 object retrieval 24
4.2 toleration of shifted proposals 26
4.3 object discovery ability 26
5 Conclusion 29

[1] D. H. Ballard. Generalizing the hough transform to detect arbitrary shapes. Pattern Recognition, 13(2):111-122, 1981.
[2] M. Cho, S. Kwak, C. Schmid, and J. Ponce. Unsupervised object discovery and localization in the wild: Part-based matching with bottom-up region proposals. In CVPR, pages 1201-1210. IEEE Computer Society, 2015.
[3] S. Chopra, R. Hadsell, and Y. LeCun. Learning a similarity metric discriminatively, with application to face verification. In CVPR (1), pages 539-546. IEEE Computer Society, 2005.
[4] O. Chum and A. Zisserman. An exemplar model for learning object classes. In CVPR. IEEE Computer Society, 2007.
[5] T. Deselaers, B. Alexe, and V. Ferrari. Localizing objects while learning their appearance. In ECCV (4), volume 6314 of Lecture Notes in Computer Science, pages 452-466. Springer, 2010.
[6] T. Deselaers, B. Alexe, and V. Ferrari. Weakly supervised localization and learning with generic knowledge. International Journal of Computer Vision, 100(3):275-293, 2012.
[7] R. Hadsell, S. Chopra, and Y. LeCun. Dimensionality reduction by learning an invariant mapping. In CVPR (2), pages 1735-1742. IEEE Computer Society, 2006.
[8] Z. Hayder, X. He, and M. Salzmann. Structural kernel learning for large scale multiclass object co-detection. In ICCV, pages 2632-2640. IEEE Computer Society, 2015.
[9] Z. Hayder, M. Salzmann, and X. He. Object co-detection via efficient inference in a fully-connected CRF. In ECCV (3), volume 8691 of Lecture Notes in Computer Science, pages 330-345. Springer, 2014.
[10] A. Joulin, K. D. Tang, and F. Li. Efficient image and video co-localization with frank-wolfe algorithm. In ECCV (6), volume 8694 of Lecture Notes in Computer Science, pages 253-268. Springer, 2014.
[11] H. Liu and S. Yan. Common visual pattern discovery via spatially coherent correspondences. In CVPR, pages 1609-1616. IEEE Computer Society, 2010.
[12] S. Ren, K. He, R. B. Girshick, and J. Sun. Faster R-CNN: towards real-time object detection with region proposal networks. In NIPS, pages 91-99, 2015.
[13] M. Rubinstein, A. Joulin, J. Kopf, and C. Liu. Unsupervised joint object discovery and segmentation in internet images. In CVPR, pages 1939-1946. IEEE Computer Society, 2013.
[14] B. C. Russell, W. T. Freeman, A. A. Efros, J. Sivic, and A. Zisserman. Using multiple segmentations to discover objects and their extent in image collections. In CVPR (2), pages 1605-1614. IEEE Computer Society, 2006.
[15] F. Schro, D. Kalenichenko, and J. Philbin. Facenet: A unied embedding for face recognition and clustering. In CVPR, pages 815-823. IEEE Computer Society, 2015.
[16] K. Simonyan and A. Zisserman. Very deep convolutional networks for large-scale image recognition. CoRR, abs/1409.1556, 2014.
[17] P. Siva and T. Xiang. Weakly supervised object detector learning with model drift detection. In ICCV, pages 343-350. IEEE Computer Society, 2011.
[18] K. D. Tang, A. Joulin, L. Li, and F. Li. Co-localization in real-world images. In CVPR, pages 1464-1471. IEEE Computer Society, 2014.
[19] Y. Wei, F. Wen, W. Zhu, and J. Sun. Geodesic saliency using background priors. In ECCV (3), volume 7574 of Lecture Notes in Computer Science, pages 29-42. Springer, 2012.
[20] K. Q. Weinberger, J. Blitzer, and L. K. Saul. Distance metric learning for large margin nearest neighbor classication. In NIPS, pages 1473-1480, 2005.
[21] C. L. Zitnick and P. Dollar. Edge boxes: Locating object proposals from edges. In Computer Vision - ECCV 2014 - 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part V, pages 391-405, 2014.