Search Results (6)

In the evolving landscape of computer vision, the integration of machine learning algorithms with cutting-edge hardware platforms is increasingly pivotal, especially in the context of disruptive healthcare systems. This study introduces an optimized implementation of a Convolutional Neural Network (CNN) on the Basys3 FPGA, designed specifically for accelerating the classification of cytotoxicity in human kidney cells. Addressing the challenges posed by constrained dataset sizes, compute-intensive AI algorithms, and hardware limitations, the approach presented in this paper leverages efficient image augmentation and pre-processing techniques to enhance both prediction accuracy and the training efficiency. The CNN, quantized to 8-bit precision and tailored for the FPGA’s resource constraints, significantly accelerates training by a factor of three while consuming only 1.33% of the power compared to a traditional software-based CNN running on an NVIDIA K80 GPU. The network architecture, composed of seven layers with excessive hyperparameters, processes downscale grayscale images, achieving notable gains in speed and energy efficiency. A cornerstone of our methodology is the emphasis on parallel processing, data type optimization, and reduced logic space usage through 8-bit integer operations. We conducted extensive image pre-processing, including histogram equalization and artefact removal, to maximize feature extraction from the augmented dataset. Achieving an accuracy of approximately 91% on unseen images, this FPGA-implemented CNN demonstrates the potential for rapid, low-power medical diagnostics within a broader IoT ecosystem where data could be assessed online. This work underscores the feasibility of deploying resource-efficient AI models in environments where traditional high-performance computing resources are unavailable, typically in healthcare settings, paving the way for and contributing to advanced computer vision techniques in embedded systems. Full article

An effective equaliser is crucial for eliminating inconsistencies in the connected serial batteries and extending the life of the battery system. The current equalisers generally have the problems of low equalisation efficiency, slow equalisation speed, and complex switching control. A layered parallel equaliser based on a flyback transformer multiplexed for a lithium-ion battery system is proposed. The equaliser employs both hierarchical and parallel equalisation techniques, allowing for simultaneous processing of multiple objectives. This enhances both the efficiency and speed of the equalisation process. The efficiency of equalisation can be further improved by implementing PWM control with deadband complement. Additionally, the flyback transformer serves as an energy storage component for both layers of the equalisation module, resulting in a significant reduction in the size and cost of the equaliser. The circuit topology of the equaliser is presented, and its operational principle, switching control, and equalisation control strategy are analysed in detail. Finally, an experimental platform consisting of six lithium-ion batteries is constructed, and equalisation experiments are conducted to verify the advantages of the proposed equaliser in terms of equalisation speed, efficiency, and cost. Full article

One of the most promising research areas in the healthcare industry and the scientific community is focusing on the AI-based applications for real medical challenges such as the building of computer-aided diagnosis (CAD) systems for breast cancer. Transfer learning is one of the recent emerging AI-based techniques that allow rapid learning progress and improve medical imaging diagnosis performance. Although deep learning classification for breast cancer has been widely covered, certain obstacles still remain to investigate the independency among the extracted high-level deep features. This work tackles two challenges that still exist when designing effective CAD systems for breast lesion classification from mammograms. The first challenge is to enrich the input information of the deep learning models by generating pseudo-colored images instead of only using the input original grayscale images. To achieve this goal two different image preprocessing techniques are parallel used: contrast-limited adaptive histogram equalization (CLAHE) and Pixel-wise intensity adjustment. The original image is preserved in the first channel, while the other two channels receive the processed images, respectively. The generated three-channel pseudo-colored images are fed directly into the input layer of the backbone CNNs to generate more powerful high-level deep features. The second challenge is to overcome the multicollinearity problem that occurs among the high correlated deep features generated from deep learning models. A new hybrid processing technique based on Logistic Regression (LR) as well as Principal Components Analysis (PCA) is presented and called LR-PCA. Such a process helps to select the significant principal components (PCs) to further use them for the classification purpose. The proposed CAD system has been examined using two different public benchmark datasets which are INbreast and mini-MAIS. The proposed CAD system could achieve the highest performance accuracies of 98.60% and 98.80% using INbreast and mini-MAIS datasets, respectively. Such a CAD system seems to be useful and reliable for breast cancer diagnosis. Full article

Voltages of electric double-layer capacitor (EDLC) modules vary rather wider than traditional secondary batteries. Although EDLCs should desirably be cycled in a voltage range as wide as possible to achieve a high energy utilization ratio, the wide voltage variation of EDLC modules impairs the performance of DC–DC converters. To address such issues, previous works reported series-parallel reconfiguration techniques, which are roughly divided into balance- and unbalance-shift circuits. However, conventional balance-shift circuits are not applicable to modules comprising odd number cells, impairing modularity. Unbalance-shift circuits, on the other hand, unavoidably cause cell voltage imbalance that reduces energy utilization ratio. This paper proposes a novel series-parallel reconfigurable EDLC module with cell voltage equalization capability. The proposed reconfigurable EDLC module is applicable to any number of cells, realizing good modularity. Furthermore, all cells in the proposed module can be charged and discharged uniformly without generating cell voltage imbalance, achieving an improved energy utilization ratio compared with conventional techniques. A five-cell module prototype was built for experimental verification. While the module voltage varied between 1.04 and 2.83 V, all cells discharged from 2.5 to 0.3 V. The result is equivalent to a 98.6% energy utilization ratio. Full article

Voltage variations of electric double-layer capacitors (EDLCs) are rather wider than those of traditional rechargeable batteries, and an energy utilization ratio of EDLCs is dependent on cells’ voltage variation ranges. To satisfactorily utilize EDLCs’ energies, voltages of EDLC modules should be within a certain range, while cells need to be charged and discharged over the wide voltage range. To this end, various kinds of series-parallel reconfiguration techniques based on balance- and unbalance-shift circuits have been proposed, but conventional techniques can only be applied to modules consisting of even number cells, impairing the design flexibility and scalability. With the unbalance-shift circuits, cell voltages are unavoidably mismatched due to unequal currents, resulting in reduced energy utilization ratios. This article proposes a novel series-parallel reconfiguration technique with voltage equalization capability for EDLC modules. The proposed technique can be applied to any number of cells, improving design flexibility and scalability. Furthermore, since the proposed circuit behaves as a switched capacitor converter, in which all cells are virtually connected in parallel, cells are equally charged and discharged without causing voltage imbalance, realizing the improved energy utilization ratio. A prototype for an EDLC module comprising four cells, each with a rated charging voltage of 2.5 V, was built and experimentally tested. The module voltage varied in the range of 3.2–5.0 V, while all cells were uniformly discharged down to as low as 0.8 V, achieving the energy utilization ratio of 90%. Full article

To eliminate the influence of the inconsistency on the cycle life and the available capacity of battery packs, and improve the balancing speed, a novel inductor-based layered bidirectional equalizer (IBLBE) is proposed. The equalizer is composed of two layers of balancing circuits connected in parallel. Each layer contains multiple balancing sub-circuits based on buck-boost converters. These balancing sub-circuits can equalize the corresponding cells simultaneously, and allow the dynamic adjustment of equalization path and equalization threshold. Analysis and simulation results demonstrate the IBLBE has a higher level balancing speed than other equalizers based on switched-capacitor or switched-inductor converters, and reduces the balancing time by 30% compared to existing inductor-based parallel architecture equalizers (PAEs). Experimental results are presented to validate the analysis and effectiveness of the proposed equalizer. Full article