BinaryCoP: Binary Neural Network-based COVID-19 Face-Mask Wear and Positioning Predictor on Edge Devices

Abstract

Face masks have long been used in many areas of everyday life to protect against the inhalation of hazardous fumes and particles. They also offer an effective solution in healthcare for bi-directional protection against air-borne diseases. Wearing and positioning the mask correctly is essential for its function. Convolutional neural networks (CNNs) offer an excellent solution for face recognition and classification of correct mask wearing and positioning. In the context of the ongoing COVID-19 pandemic, such algorithms can be used at entrances to corporate buildings, airports, shopping areas, and other indoor locations, to mitigate the spread of the virus. These application scenarios impose major challenges to the underlying compute platform. The inference hardware must be cheap, small and energy efficient, while providing sufficient memory and compute power to execute accurate CNNs at a reasonably low latency. To maintain data privacy of the public, all processing must remain on the edge-device, without any communication with cloud servers. To address these challenges, we present BinaryCoP, a low-power binary neural network classifier for correct facial-mask wear and positioning. The classification task is implemented on an embedded FPGA accelerator, performing high-throughput binary operations. Classification can take place at up to ∼6400 framesper-second, easily enabling multi-camera, speed-gate settings or statistics collection in crowd settings. When deployed on a single entrance or gate, the idle power consumption is reduced to 1.6W, improving the battery-life of the device. We achieve an accuracy of up to 98% for four wearing positions of the MaskedFace-Net dataset. To maintain equivalent classification accuracy for all face structures, skin-tones, hair types, and mask types, the algorithms are tested for their ability to generalize the relevant features over all subjects using the Grad-CAM approach.

GAAF: Searching Activation Functions for Binary Neural Networks Through Genetic Algorithm

“…Consequently, the costly floating-point dot-product (or multiply accumulation) computation can be replaced by lighgtweight bit-wise exclusive-nor (xnor) and population-count (popc) operations. BNNs show great potential in cost-and power-restricted domains [3][4][5][6][7] , such as Internet-of-things (IoT), embedded systems, mobile devices, etc. This is primarily due to the significantly reduced computation requirements, lower power cost, lower storage demand, and improved robustness to input noise [1,8,9] .…”

“…Existing research has shown that, through BNN, significant speedups and memory reduction can be harvested on CPUs [10] , GPUs [8,11] , and FPGAs [12][13][14] , compared with their DNN counterparts. As a result, BNNs have been adopted in a variety of practical applications such as COVID-19 face-cover detection [3] , auto-driving [4] , smart agriculture [5] , image enhancement [6] , 3D objection detection [7] , etc.…”

GAAF: Searching Activation Functions for Binary Neural Networks through Genetic Algorithm

“…Consequently, the costly floatingpoint dot-product (or multiply accumulation) computation can be replaced by lighgtweight bit-wise exclusive-nor (xnor) and population-count (popc) operations. BNNs show great potential in cost-& power-restricted domains [3]- [7], such as Internet-of-Things (IoT), embedded systems, mobile devices, etc. This is primarily due to the significantly reduced computation requirements, lower power cost, lower storage demand, and improved robustness to input noise [1], [8], [9].…”

“…Existing research has shown that, through BNN, significant speedups and memory reduction can be harvested on CPUs [10], GPUs [8], [11], and FPGAs [12]- [14], compared with their DNN counterparts. As a result, BNNs have been adopted in a variety of practical applications such as COVID-19 facecover detection [3], auto-driving [4], smart agriculture [5], image enhancement [6], 3D objection detection [7], etc.…”

Searching Similarity Measure for Binarized Neural Networks

“…applications; (iv) Power efficiency: due to simplified hardware and less memory demand, BNNs can run on extremely energyconstraint edge devices; Last but not the least, (v) Robustness: it has been recently shown that, due to the discrete nature through binarization, compared to normal DNNs, BNNs show improved robustness against noise and poison attacks [13]. Because of these features, BNNs have been showcased on a variety of practical edge applications, such as COVID-19 facecover detection [14], auto-driving [15], smart agriculture [16], image enhancement [17], 3D objection detection [18], etc.…”