Stereoscopic Image Super-Resolution with Interactive Memory Learning

Zhu, Xiangyuan; Guo, Kehua; Fang, Hui; Wu, Zheng; Tan, Xuyang; Liu, Chao

doi:10.2139/ssrn.4308294

Cited by 2 publications

(10 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, we consider {4, 6, 8, 16} precisions in our search space for the ImageNet dataset as the accelerator supports these bit widths. Hence, there exist 16 distinct weight and activation precision choices for each layer, which are as follows: {(4,4), (6,4), (8,4), (16,4), (4,6), (6,6), (8,6), (16,6), (4,8), (6,8), (8,8), (16,8), (4,16), (6,16), (8,16), (16,16)}…”

Section: ) 1×1 Standard Spatial Convolutionmentioning

confidence: 99%

“…In this paper, we consider 0.001 as the grouping latency g l for MobilenetV2 on Bitfusion. The 16 paths in layer 12 are grouped, resulting in the following clusters: (i) {(4,4), (6,4), (8,4), (16,4)}, (ii) {(4,6), (6,6), (8,6), (16,6)}, (iii) {(4,8), (6,8), (8,8), (16,8)}, (iv) {(4,16), (6,16), (8,16), (16,16)}, as shown in Figure 3.…”

Section: ) 1×1 Standard Spatial Convolutionmentioning

confidence: 99%

“…The Supernetwork possesses the same dimensions (#layers, kernel and channel size) as the given pretrained network. (6,4), (8,8), (6,6), (16,4), (4,8)}. The five architectural parameters {α 1 , α 2 , α 3 , α 4 , α 5 } are assigned to each of the choices in the search space {(W, A)} = {(6,4), (8,8), (6,6), (16,4), (4,8)}.…”

Section: A Supernetwork Constructionmentioning

confidence: 99%

“…(6,4), (8,8), (6,6), (16,4), (4,8)}. The five architectural parameters {α 1 , α 2 , α 3 , α 4 , α 5 } are assigned to each of the choices in the search space {(W, A)} = {(6,4), (8,8), (6,6), (16,4), (4,8)}. Each architectural parameter (α i ) is a Softmax representation over all the architectural parameters in the same layer, as shown in Equation 4.…”

Section: A Supernetwork Constructionmentioning

confidence: 99%

“…Convolutional Neural Networks (CNNs) have attained remarkable results on a wide range of tasks, such as Image Classification [1], Super Resolution [2]- [4], etc., and are often being deployed in various real-time scenarios on the cloud and edge platforms. This success is mainly attributed to efficient neural network architecture design in terms of the kernel and channel sizes, methods for efficient inference methods such as Pruning [5], [6] and Quantization [7], [8], and systematic domain-specific hardware configuration for network acceleration [9], [10].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Differentiable Neural Architecture, Mixed Precision and Accelerator Co-Search

Chitty-Venkata,

Bian,

Emani

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Quantization, effective Neural Network architecture, and efficient accelerator hardware are three important design paradigms to maximize accuracy and efficiency. Mixed Precision Quantization is a process of assigning different precision to different Neural Network layers for optimized inference. Neural Architecture Search (NAS) is a process of automatically designing the neural network for a task and can also be extended to search for the precision of each weight and activation matrix. In this paper, we develop the following three methods: (i) Fast Differentiable Hardware-aware Mixed Precision Quantization Search method to find optimal precision, (ii) Joint Differentiable hardware-aware Architecture and Mixed Precision Quantization Co-search, (iii) Joint Accelerator, Architecture, and Precision triple co-search to find best possibilities in all the three worlds. We demonstrate the effectiveness of our proposed methods targeting Bitfusion accelerator by searching mixed precision models on MobilenetV2. We achieve better accuracy-latency trade-off models than the manually designed and previously proposed search methods.

show abstract

Section: ) 1×1 Standard Spatial Convolutionmentioning

confidence: 99%

Section: ) 1×1 Standard Spatial Convolutionmentioning

confidence: 99%

Section: A Supernetwork Constructionmentioning

confidence: 99%

Section: A Supernetwork Constructionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Differentiable Neural Architecture, Mixed Precision and Accelerator Co-Search

Chitty-Venkata,

Bian,

Emani

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

Classification of Lung and Colon Cancer Histopathological Images Using Global Context Attention Based Convolutional Neural Network

Provath,

Deb,

Dhar

et al. 2023

IEEE Access

View full text Add to dashboard Cite

The malignant neoplastic malady known as cancer appears to exhibit a significantly elevated rate of mortality owing to its virulence and pronounced propensity for metastasis. To augment the diagnostic efficacy, research endeavors have been undertaken utilizing complex deep learning architectures. However, the performance of these efforts remains circumscribed by smaller dataset size, quality of the data, the interclass variations present between lung adenocarcinoma and lung squamous cell carcinoma, and the complexity of deploying to mobile devices and failure to address both image and patient level accuracy measurements. To surmount these obstacles, the present study proposes a stage-based method for enhancing the images, in conjunction with utilizing a global context attention-guided convolutional neural network that effectively captures both channel and spatial information and semantic information extracted from the input image. Implementing the proposed methodology increased total image level accuracy to 99.76% and a patient level accuracy of 96.5%, a metric that has yet to be previously quantified. The addition of the global context attention module decreases the model's parameter count by 0.47 million, reduces the computational costs by saving 10.54 million floating point operations per second (FLOPs) and 10.72 million multiply-accumulate operations (MACs), and results in a 0.03s improvement in inference time. Furthermore, this module enhances both image level and patient level accuracy, boosting them by 2.84% and 3.17%, respectively, compared to using only the convolutional block attention module in the baseline convolutional neural network. Consequently, this modification renders the model highly suitable for deployment on mobile devices due to its adaptability. Our findings provide supporting evidence for the potential of this method to serve as a noninvasive screening tool capable of reliably classifying lung and colon cancer subtypes.INDEX TERMS Stage-based method, global context attention module, convolutional neural network, noninvasive screening tool, lung and colon cancer.

show abstract

Stereoscopic Image Super-Resolution with Interactive Memory Learning

Cited by 2 publications

References 0 publications

Differentiable Neural Architecture, Mixed Precision and Accelerator Co-Search

Differentiable Neural Architecture, Mixed Precision and Accelerator Co-Search

Classification of Lung and Colon Cancer Histopathological Images Using Global Context Attention Based Convolutional Neural Network

Contact Info

Product

Resources

About