An FPGA-based Hybrid Memory Emulation System

Wen, Fei; Qin, Mian; Gratz, Paul V.; Reddy, Narasimha

doi:10.1109/fpl53798.2021.00039

Cited by 8 publications

(2 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Mouth recognition, like the recognition of other human body parts, is essential in the image-based biomedical analysis of human features. The recognition methods can be generally classified into two categories: 1) unsupervised or semi-supervised feature deep learning algorithms [1]; 2) traditional pattern recognition algorithms [2]. In general, deep learning algorithms have high accuracy.…”

Section: Introductionmentioning

confidence: 99%

A Pipeline-based Low Complexity Dual Thresholds and Multiresolution Mouth Detection Method

Lin

Lan

et al. 2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

Mouth detection is a basic step in the automatic diagnosis of buccal diseases. This paper proposes a method of using FPGA(Field Prog ram mable Gate Array) to detect the human mouth region in images, which adopts a pipelined design for high-speed operating. Thus, it can be applied to mouth motion detection as well. After an original image is processed regularly by median filtering, color conversion is performed to obtain its corresponding H component, which is used to create a partitioned binary image. Each constituent region of the binary image uses two thresholds for regional processing separately, which results in two binary images: one comprises the mouth and some other regions with a color close to that of the mouth, and the other indicates the position of the mouth. A logical AND is performed on these two binary images to calculate the border of mouth region with reduced resolution, which is then superimposed on the buffered original image to confirm the mouth region eventually. An FPGA-based scheme is presented, which shows the method could detect the whole mouth properly, with a maximum clock speed of 150MHz. In addition, processing an image of 200*200 only consumes 2178 Slice LUTs(Look Up Table), which is a relatively low resource consumption.

show abstract

Section: Introductionmentioning

confidence: 99%

A Pipeline-based Low Complexity Dual Thresholds and Multiresolution Mouth Detection Method

Lin

Lan

et al. 2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

show abstract

“…Short term monitoring cannot guarantee capturing participants' symptoms and their associated behavior changes when they are unwell. Meanwhile, if we retrain models using all existing data, the computation and memory burden would be high and prohibitive for mobile devices [11][12][13]. If we retrain models using only the newly incoming data, a classic problem referred to as catastrophic forgetting will significantly downgrade model performance.…”

Section: Introductionmentioning

confidence: 99%

Semi-supervised Graph Instance Transformer for Mental Health Inference

Dong

Tang

Cai

et al. 2021

2021 20th IEEE International Conference on Machine Learning and Applications (ICMLA)

View full text Add to dashboard Cite

Mobile sensing appears as a promising solution for health inference problem (e.g., influenza-like symptom recognition) by leveraging diverse smart sensors to capture fine-grained information about human behaviors and ambient contexts. Centralized training of machine learning models can place mobile users' sensitive information under privacy risks due to data breach and misexploitation. Federated Learning (FL) enables mobile devices to collaboratively learn global models without the exposure of local private data. However, there are challenges of on-device FL deployment using mobile sensing: 1) long-term and continuously collected mobile sensing data may exhibit domain shifts as sensing objects (e.g. humans) have varying behaviors as a result of internal and/or external stimulus; 2) model retraining using all available data may increase computation and memory burden; and 3) the sparsity of annotated crowd-sourced data causes supervised FL to lack robustness. In this work, we propose FedMobile, an incremental semi-supervised federated learning algorithm, to train models semi-supervisedly and incrementally in a decentralized online fashion. We evaluate FedMobile using a real-world mobile sensing dataset for influenza-like symptom recognition. Our empirical results show that FedMobile-trained models achieve the best results in comparison to the selected baseline methods.

show abstract