Search

This paper presents a supervised contrastive learning (SCL) framework for respiratory sound classification and the hardware implementation of learned ResNet on field programmable gate array (FPGA) for real-time monitoring. At the algorithmic level, multiple techniques such as features augmentation and MixUp are combined holistically to mitigate the impact of data scarcity and imbalanced classes in the training dataset. Bayesian optimization further enhances the classification accuracy through parameter tuning in pre-processing and SCL. The proposed framework achieves 0.8725 total score (including runtime score) on a ResNet-18 model in both event and record multi-class classification tasks using the SJTU Paediatric Respiratory Sound Database (SPRSound). In addition, algorithm-hardware co-optimizations including Quantization-Aware Training (QAT), merge of network layers, optimization of memory size and number of parallel threads are performed for hardware implementation on FPGA. This approach reduces 40% model size and 70% computation latency. The learned ResNet is implemented on a Xilinx Zynq ZCU102 FPGA with 16ms latency and less than 2% inference score degradation compared to the software model.

Bioretention systems have been implemented as stormwater best management practices (BMPs) worldwide to treat non-point sources pollution. Due to insufficient research, the design guidelines for bioretention systems in tropical countries are modeled after those of temperate countries. However, climatic factors and stormwater runoff characteristics are the two key factors affecting the capacity of bioretention system. This paper reviews and compares the stormwater runoff characteristics, bioretention components, pollutant removal requirements, and applications of bioretention systems in temperate and tropical countries. Suggestions are given for bioretention components in the tropics, including elimination of mulch layer and submerged zone. More research is required to identify suitable additives for filter media, study tropical shrubs application while avoiding using grass and sedges, explore function of soil faunas, and adopt final discharged pollutants concentration (mg/L) on top of percentage removal (%) in bioretention design guidelines.

In August 1994, blood samples were collected from 110 pigs at a slaughter house in Bandar Lampung (Sumatra) and from 98 pigs in a slaughter house in Ujung Pandang (Sulawesi), 208 pigs in total. Anti-Toxoplasma gondii antibody titer in these samples was surveyed using the latex agglutination test. The positivity rate was 3.6% (4/110) in Bandar Lampung, and 9.2% (9/98) in Ujung Pandang. The antibody titer was 1:128 in two and 1:512 in two in Bandar Lampung, and 1:64 in six, 1:128 in two, and 1:256 in one in Ujung Pandang, showing that the titer was higher in Bandar Lampung.