Instance Segmentation Network With Self-Distillation for Scene Text Detection

Yang, Peng; Yang, Guowei; Gong, Xun; Wu, Pingping; Han, Xu; Wu, Jiasong; Chen, Caisen

doi:10.1109/access.2020.2978225

Cited by 13 publications

(7 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The PSENet [27] uses convolutional features extracted with FPN [28] and an iterative framework which grows text regions from their innermost pixels allowing the network to accurately separate the individual text region instances. Another idea is to combine region level detection with mask coefficients that map pixels to specific text regions [29]. In this paper, we propose a model that predicts a pixel-level mask of text regions.…”

Section: B Scene Text Detectionmentioning

confidence: 99%

FCN-LectureNet: Extractive Summarization of Whiteboard and Chalkboard Lecture Videos

Davila

Setlur

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Recording and sharing of educational or lecture videos has increased in recent years. Within these recordings, we find a large number of math-oriented lectures and tutorials which attract students of all levels. Many of the topics covered by these recordings are better explained using handwritten content on whiteboards or chalkboards. Hence, we find large numbers of lecture videos that feature the instructor writing on a surface. In this work, we propose a novel method for extraction and summarization of the handwritten content found in such videos. Our method is based on a fully convolutional network, FCN-LectureNet, which can extract the handwritten content from the video as binary images. These are further analyzed to identify the unique and stable units of content to produce a spatial-temporal index of handwritten content. A signal which approximates content deletion events is then built using information from the spatial-temporal index. The peaks of this signal are used to create temporal segments of the lecture based on the notion that sub-topics change when large portions of content are deleted. Finally, we use these segments to create an extractive summary of the handwritten content based on key-frames. This will facilitate content-based search and retrieval of these lecture videos. In this work, we also extend the AccessMath dataset to create a novel dataset for benchmarking of lecture video summarization called LectureMath. Our experiments on both datasets show that our novel method can outperform existing methods especially on the larger and more challenging dataset. Our code and data will be made publicly available.INDEX TERMS fully convolutional networks, handwritten text detection, image binarization, lecture videos, video summarization.

show abstract

Section: B Scene Text Detectionmentioning

confidence: 99%

FCN-LectureNet: Extractive Summarization of Whiteboard and Chalkboard Lecture Videos

Davila

Setlur

et al. 2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Yang et al [74] utilized the information of earlier training epochs to supervise the later training epochs. The self-distillation mechanism has been applied in many fields like classification [80], weakly-supervised object detection [28], text segmentation [75], etc. In this paper, we introduce the self-distillation mechanism into the training process of the WSSS model and propose a novel self-dual teaching strategy to facilitate an effective knowledge distillation under the weak supervision.…”

Section: Related Workmentioning

confidence: 99%

Weakly Supervised Semantic Segmentation via Alternative Self-Dual Teaching

Zhang¹,

Zeng²,

Guo³

et al. 2021

Preprint

View full text Add to dashboard Cite

Current weakly supervised semantic segmentation (WSSS) frameworks usually contain the separated mask-refinement model and the main semantic region mining model. These approaches would contain redundant feature extraction backbones and biased learning objectives, making them computational complex yet sub-optimal to addressing the WSSS task. To solve this problem, this paper establishes a compact learning framework that embeds the classification and mask-refinement components into a unified deep model. With the shared feature extraction backbone, our model is able to facilitate knowledge sharing between the two components while preserving a low computational complexity. To encourage high-quality knowledge interaction, we propose a novel alternative self-dual teaching (ASDT) mechanism. Unlike the conventional distillation strategy, the knowledge of the two teacher branches in our model is alternatively distilled to the student branch by a Pulse Width Modulation (PWM), which generates PW wave-like selection signal to guide the knowledge distillation process. In this way, the student branch can help prevent the model from falling into local minimum solutions caused by the imperfect knowledge provided of either teacher branch. Comprehensive experiments on the PASCAL VOC 2012 and COCO-Stuff 10K demonstrate the effectiveness of the proposed alternative self-dual teaching mechanism as well as the new state-of-the-art performance of our approach.

show abstract

“…By this way, the students perform significantly better than the teacher in language modeling tasks. Hence, Yang et al [56] utilized the self-distillation to accurately detect the text in the image and optimized the teacher-student training process. Clark et al [57] applied the BAN to multitask learning and validated its effectiveness in other NLP tasks such as textual similarity, textual entailment, and so on.…”

Section: B Knowledge Distillationmentioning

confidence: 99%

A Mutually Auxiliary Multitask Model With Self-Distillation for Emotion-Cause Pair Extraction

Liu

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Instance Segmentation Network With Self-Distillation for Scene Text Detection

Cited by 13 publications

References 39 publications

FCN-LectureNet: Extractive Summarization of Whiteboard and Chalkboard Lecture Videos

FCN-LectureNet: Extractive Summarization of Whiteboard and Chalkboard Lecture Videos

Weakly Supervised Semantic Segmentation via Alternative Self-Dual Teaching

A Mutually Auxiliary Multitask Model With Self-Distillation for Emotion-Cause Pair Extraction

Contact Info

Product

Resources

About