CNN Inference acceleration using low-power devices for human monitoring and security scenarios

Abstract

Security is currently one of the top concerns in our society. From governmental installations to private companies and medical institutions, they all have to address directly with security issues as: access to restricted information quarantine control, or criminal tracking. As an example, identifying patients is critical in hospitals or geriatrics in order to isolate infected people, which has proven to be a non- trivial issue with the COVID-19 pandemic that is currently affecting all countries, or to locate fled patients. Face recognition is then a non-intrusive alternative for performing these tasks. Although FaceNet from Google has proved to be almost perfect, in a multi-face scenario its performance decays rapidly. In order to mitigate this loss of performance, in this paper a cluster based on the Neural Computer Stick version 2 and OpenVINO by Intel is proposed. A detailed power and runtime study is shown for two programming models, namely: multithreading and multiprocessing. Furthermore, 3 different hosts have been considered. In the most efficient configuration, an average of 6 frames per second has been achieved using the Raspberry Pi 4 as host and with a power consumption of just 11.2W, increasing by a factor of 3.3X the energy efficiency with respect to a PC-based solution in a multi-face scenario.

An ensemble multi-stream classifier for infant needs detection

“…As mentioned above, the majority of research in this field was conducted on adults. Numerous researchers focused on emotion detection from images [10] , [11] , [12] , [13] , [14] , [15] , [16] , [17] , [18] , [19] , [20] . Emopy, a human sentiment classifier, is also a good approach [21] .…”

Continuous Monitoring of Elastic Modulus of Mortars Using a Single-Board Computer and Cost-Effective Components

“…Multithreading is a technique that allows a single process to have multiple threads of execution. Each thread has its own program counter, stack, and local variables, but shares the same memory space as the other threads within the same process [46]. This It is important to note that multiprocessing must be differentiated from multithreading, which is a popular programming technique that is more lightweight than multiprocessing.…”

“…Multithreading is a technique that allows a single process to have multiple threads of execution. Each thread has its own program counter, stack, and local variables, but shares the same memory space as the other threads within the same process [46]. This allows for the concurrent execution of multiple parts of a program, (threads) within a single process, for compatible devices and compatible programming environments.…”

“…This lock was implemented to safeguard the access to the resources of the Python interpreter and allow only one thread at a time to acquire the lock. This approach improves the efficiency of single-core environments, as the GIL can be temporarily unlocked when a thread is waiting for an I/O operation, thereby enabling time slicing techniques to optimize the performance [46]. Consequently, the GIL can limit the performance benefits of multi-threaded programs on multi-core systems, as only one thread can run at a time.…”

A Deep Neural Network for Cervical Cell Classification Based on Cytology Images

“…For example, [16] develops a hardware-efficient dataflow based on CNN to reduce computations. Reference [17] proposes a low-power and low-cost NCS2 cluster to accelerate the inference stage. Reference [18] constructs a low-cost real-time hardware accelerator and decreases extra internal memory usage.…”