Efficient Dense Modules of Asymmetric Convolution for Real-Time Semantic Segmentation

Lo, Shao-Yuan; Hang, Hsueh-Ming; Chan, Sheng-Wei; Lin, Jing-Jhih

doi:10.1145/3338533.3366558

Cited by 157 publications

(91 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since the purpose of ablation experiments here is to verify our approach, training the models through 2/3 number of iterations compared to the implementation of the authors of EDANet is enough. Hence, the mIoU accuracy of EDANet-RGB is slightly lower than that reported in the EDANet paper [16]. Next, EDANet-DCT is obviously less accurate than EDANet-RGB though they have identical computational cost and an identical amount of input information.…”

Section: Resultsmentioning

confidence: 72%

“…In this section, we first describe the Frequency Component Rearrangement (FCR) technique, which is used to rearrange the DCT coefficients to make them be easily exploited by CNNs. Next, we introduce the proposed DCT-EDANet, a modification of EDANet [16] to operate in the DCT domain.…”

Section: Methodsmentioning

confidence: 99%

“…We do not include any testing tricks, such as multi-crop and multiscale testing. In our experiments, we train models with one-stage and through just 2/3 number of iterations compared to that in EDANet [16] since we compare the relative accuracy in our analysis. Our experiments are conducted on a single GTX 1080Ti.…”

Section: Benchmark and Training Proceduresmentioning

confidence: 99%

“…Deep convolutional neural networks (CNNs) have shown tremendous success in a variety of computer vision tasks, such as image classification [10,13], object detection [6,15], and semantic segmentation [1,16,17]. These CNN models have been broadly adopted not only in academia but also in industry.…”

Section: Introductionmentioning

confidence: 99%

“…We call this technique as Frequency Component Rearrangement (FCR). Next, we choose EDANet [16] as our baseline network owing to its good balance between performance and complexity. Because the spatial resolution of the FCRed DCT representation is 1/8 to its RGB counterpart, the three downsampling operations of EDANet would make the feature size too small.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Exploring Semantic Segmentation on the DCT Representation

Hang

2019

Proceedings of the ACM Multimedia Asia

Self Cite

View full text Add to dashboard Cite

Typical convolutional networks are trained and conducted on RGB images. However, images are often compressed for memory savings and efficient transmission in real-world applications. In this paper, we explore methods for performing semantic segmentation on the discrete cosine transform (DCT) representation defined by the JPEG standard. We first rearrange the DCT coefficients to form a preferred input type, then we tailor an existing network to the DCT inputs. The proposed method has an accuracy close to the RGB model at about the same network complexity. Moreover, we investigate the impact of selecting different DCT components on segmentation performance. With a proper selection, one can achieve the same level accuracy using only 36% of the DCT coefficients. We further show the robustness of our method under quantization errors. To our knowledge, this paper is the first to explore semantic segmentation on the DCT representation.

show abstract

Section: Resultsmentioning

confidence: 72%

Section: Methodsmentioning

confidence: 99%

Section: Benchmark and Training Proceduresmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Exploring Semantic Segmentation on the DCT Representation

Hang

2019

Proceedings of the ACM Multimedia Asia

Self Cite

View full text Add to dashboard Cite

show abstract

ASFNet: Adaptive multiscale segmentation fusion network for real‐time semantic segmentation

Zha

Liu

Yang

et al. 2021

Computer Animation & Virtual

View full text Add to dashboard Cite

Recently, the development of deep learning has facilitated continuous progress in the field of computer vision. Pixel-level semantic segmentation serves as a fundamental task in computer vision. It achieves significant results by connecting wider and deeper backbone networks and building fine-grained segmentation heads. However, applications such as self-driving cars are more critical to the computational speed of the algorithms. The trade-off between accuracy and real-time performance of existing algorithms is still a challenging task. To address this challenge, this article proposes an adaptive multiscale segmentation fusion network to fuse multiscale contextual, which designs an adaptive multiscale segmentation fusion module based on an attention mechanism. Using segmentation fusion instead of feature fusion, the multiscale segmentation results are aggregated to obtain more precise segmentation results. The final results achieved 70.9% mIoU of accuracy in the Cityspace test set, processing images at 61 FPS when the input is 1024 × 2048. In addition, when adjusting the input size to 512 × 1024, the images are processed at 185 FPS.

show abstract

SCU‐Net: A deep learning method for segmentation and quantification of breast arterial calcifications on mammograms

et al. 2021

View full text Add to dashboard Cite

Purpose: Measurements of breast arterial calcifications (BAC) can offer a personalized, noninvasive approach to risk-stratify women for cardiovascular diseases such as heart attack and stroke. We aim to detect and segment breast arterial calcifications in mammograms accurately and suggest novel measurements to quantify detected BAC for future clinical applications. Methods: To separate BAC in mammograms, we propose a lightweight fine vessel segmentation method Simple Context U-Net (SCU-Net). Due to the large image size of mammograms, we adopt a patch-based way to train SCU-Net and obtain the final whole-image-size results by stitching patchwise results together. To further quantify calcifications, we test five quantitative metrics to inspect the progression of BAC for subjects: sum of mask probability metric (PM), sum of mask area metric (AM), sum of mask intensity metric (SIM), sum of mask area with threshold intensity metric TAM X ð Þ, and sum of mask intensity with threshold X metric TSIM X ð Þ . Finally, we demonstrate the ability of the metrics to longitudinally measure calcifications in a group of 26 subjects and evaluate our quantification metrics compared with calcified voxels and calcium mass on breast CT for 10 subjects. Results: Our segmentation results are compared with state-of-the-art network architectures based on recall, precision, accuracy, F1 score/Dice score, and Jaccard index evaluation metrics and achieve corresponding values of 0.789, 0.708, 0.997, 0.729, and 0.581 for whole-image-size results. The quantification results all show >95% correlation between quantification measures on predicted masks of SCU-Net as compared to the groundtruth and measurement of calcification on breast CT. For the calcification quantification measurement, our calcification volume (voxels) results yield R 2 -correlation values of 0.834, 0.843, 0.832, 0.798, and 0.800 for the PM, AM, SIM, TAM 100 , TSIM 100 metrics, respectively; our calcium mass results yield comparable R 2 -correlation values of 0.866, 0.873, 0.840, 0.774, and 0.798 for the same metrics. Conclusions: Simple Context U-Net is a simple method to accurately segment arterial calcification retrospectively on routine mammograms. Quantification of the calcifications based on this segmentation in the retrospective cohort study has sufficient sensitivity to detect the normal progression over time and should be useful for future research and clinical applications.

show abstract

Efficient Dense Modules of Asymmetric Convolution for Real-Time Semantic Segmentation

Cited by 157 publications

References 19 publications

Exploring Semantic Segmentation on the DCT Representation

Exploring Semantic Segmentation on the DCT Representation

ASFNet: Adaptive multiscale segmentation fusion network for real‐time semantic segmentation

SCU‐Net: A deep learning method for segmentation and quantification of breast arterial calcifications on mammograms

Contact Info

Product

Resources

About