Multiple instance deep learning for weakly-supervised visual object tracking

Huang, Kaining; Shi, Yan; Zhao, Fuqi; Zhang, Zijun; Tu, Shanshan

doi:10.1016/j.image.2020.115807

Cited by 7 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a rule, m = l K l ≪ n, thus (16) gives us an opportunity of saving memory by storing (n × m) sparse matrix instead of full (n × n) co-association matrix.…”

Section: Co-association Matrix Of Cluster Ensemblementioning

confidence: 99%

See 1 more Smart Citation

Solving weakly supervised regression problem using low-rank manifold regularization

Berikov,

Litvinenko

2021

Preprint

View full text Add to dashboard Cite

We solve a weakly supervised regression problem. Under "weakly" we understand that for some training points the labels are known, for some unknown, and for others uncertain due to the presence of random noise or other reasons such as lack of resources. The solution process requires to optimize a certain objective function (the loss function), which combines manifold regularization and low-rank matrix decomposition techniques. These low-rank approximations allow us to speed up all matrix calculations and reduce storage requirements. This is especially crucial for large datasets. Ensemble clustering is used for obtaining the co-association matrix, which we consider as the similarity matrix. The utilization of these techniques allows us to increase the quality and stability of the solution. In the numerical section, we applied the suggested method to artificial and real datasets using Monte-Carlo modeling.

show abstract

“…As a rule, m = l K l ≪ n, thus (16) gives us an opportunity of saving memory by storing (n × m) sparse matrix instead of full (n × n) co-association matrix.…”

Section: Co-association Matrix Of Cluster Ensemblementioning

confidence: 99%

“…Methods based on the cluster assumption and the manifold assumption are also used [2,16]. In [12], the upper bounds of the labels' noise characteristics are obtained, and an algorithm for estimating the degree of noise is proposed using preliminary data partitioning into clusters.…”

Section: Introductionmentioning

confidence: 99%

Solving weakly supervised regression problem using low-rank manifold regularization

Berikov,

Litvinenko

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…These methods assumed for tracking a single object, and thus these are short to our problem. Several methods have been proposed for multi-object tracking [29,16]. Nwoye et al [29] proposed a weakly-supervised tracking of surgical tools appearing in endoscopic video.…”

Section: Related Workmentioning

confidence: 99%

“…The weak label in this case is a class label of the tool type, and they assumed that one tool appears for each tool type even though several different types of tools appear at a frame. Huang et al [16] tackled a similar problem of semantic object tracking. He et al [15] proposed an unsupervised trackingby-animation framework.…”

Section: Related Workmentioning

confidence: 99%

Weakly-Supervised Cell Tracking via Backward-and-Forward Propagation

Nishimura

Hayashida

Wang

et al. 2020

Preprint

View full text Add to dashboard Cite

We propose a weakly-supervised cell tracking method that can train a convolutional neural network (CNN) by using only the annotation of "cell detection" (i.e., the coordinates of cell positions) without association information, in which cell positions can be easily obtained by nuclear staining. First, we train co-detection CNN that detects cells in successive frames by using weak-labels. Our key assumption is that co-detection CNN implicitly learns association in addition to detection. To obtain the association, we propose a backward-andforward propagation method that analyzes the correspondence of cell positions in the detection maps output of co-detection CNN. Experiments demonstrated that the proposed method can associate cells by analyzing co-detection CNN. Even though the method uses only weak supervision, the performance of our method was almost the same as the state-of-the-art supervised method. Code is publicly available in https://github.com/naivete5656/WSCTBFP .

show abstract