Garbage classification is the first step in waste segregation, recycling, or reuse. MobileNet was used to generate a model that classifies common trash according to the following categories: glass, paper, cardboard, plastic, metal, and other trash. A dataset of 2527 trash images in .jpg extension was used for the training. The model used transfer learning from a model trained on the ImageNet Large Visual Recognition Challenge dataset. The TensorFlow for Poets git repository was cloned as a working directory to retrain the MobileNet model in 500 steps. The resulting baseline model, with a final test accuracy of 87.2% was optimized and quantized. In the Andoid app development, the optimized model (with 89.34% confidence) is preferred over the quantized model (with 1.47% confidence) based on the test using a plastic image. The model app was successfully installed in a Samsung Galaxy S6 Edge+ mobile phone. The installed mobile app successfully identified a cardboard material in an image with a cardboard container. It is recommended to rerun the training using more steps as this may improve the quantized model performance since a quantized model is fit for mobile devices than models with no quantization.

In this study, image processing and artificial neural network was used to efficiently identify the nutrients and pH level of soil with the use of Soil Test Kit (STK) and Rapid Soil Testing (RST) of the Bureau of Soils and Water Management: (1) pH, (2) Nitrogen, (3) Phosphorus, (4) Potassium, (5) Zinc, (6) Calcium, and (7) Magnesium. The composition of the system is made of five sections namely soil testing, image capturing, image processing, training system for neural network, and result. The use of Artificial Neural Network is to hasten the performance of image processing in giving accurate result. The system will base on captured image data, 70% for training, 15% for testing and 15% for validation as default of neural network tool of MATLAB. Based on the result, the program will show the qualitative level of soil nutrients and pH. Overall, this study identifies the soil nutrient and pH level of the soil and was proven accurate.

Incorrect diagnosis is still apparent especially in respiratory diseases. There is truly a need to extend the study with correct diagnosis as most of lung diseases affect children. Over-diagnosis is also a problem that is necessary to address. As an aid to health diagnostics and health professionals, this study constructed a low-cost diagnostic tool that classifies a chest x-ray image if it is under the normal or pneumonia category. Training, validation and cross-entropy were done by using MobileNetV2 as a pre-trained model and served as the general convolutional neural network system. Results yielded high accuracy based on percentage accuracy. Further validation is incorporated by showing the confusion matrix of the system.

Identifying the plant's developmental growth stages from seed leaf is crucial to understand plant science and cultivation management deeply. An efficient vision-based system for plant growth monitoring entails optimum segmentation and classification algorithms. This study presents coupled color-based superpixels and multifold watershed transformation in segmenting lettuce plant from complicated background taken from smart farm aquaponic system, and machine learning models used to classify lettuce plant growth as vegetative, head development and for harvest based on phytomorphological profile. Morphological computations were employed by feature extraction of the number of leaves, biomass area and perimeter, convex area, convex hull area and perimeter, major and minor axis lengths of the major axis length the dominant leaf, and length of plant skeleton. Phytomorphological variations of biomass compactness, convexity, solidity, plant skeleton, and perimeter ratio were included as inputs of the classification network. The extracted Lab color space information from the training image set undergoes superpixels overlaying with 1,000 superpixel regions employing K-means clustering on each pixel class. Six-level watershed transformation with distance transformation and minima imposition was employed to segment the lettuce plant from other pixel objects. The accuracy of correctly classifying the vegetative, head development, and harvest growth stages are 88.89%, 86.67%, and 79.63%, respectively. The experiment shows that the test accuracy rates of machine learning models were recorded as 60% for LDA, 85% for ANN, and 88.33% for QSVM. Comparative analysis showed that QSVM bested the performance of optimized LDA and ANN in classifying lettuce growth stages. This research developed a seamless model in segmenting vegetation pixels, and predicting lettuce growth stage is essential for plant computational phenotyping and agricultural practice optimization.

Implementing computer vision on traffic scenarios are one of the most widely sought area in the field of vision research. In dealing with the surveillance in traffic scenarios, every vehicle in the scene must be observed which results to problem arising from instances whenever the traffic density in an area is high due to occlusion caused by the large number of vehicles being observed. Thus, this paper proposes a vehicle detection and tracking algorithm whose main purpose is to detect and track vehicles entering an intersection and track them robustly in real-time. The algorithm which was used is a blob analysis and tracking based on a mean-shift kernel. The blob approach acts as the main tracking and will use the mean-shift in the event of blob merging or occlusion. In this paper, the proposed tracking method is tested using a CCTV camera on an intersection with high traffic density to illustrate the capability of solving occlusion and observe the robustness of the algorithm in the scene. The results show that the proposed system successfully tracks the vehicles during and after occlusion with other vehicles or other types of objects in the scene.

Developing countries face the problem of crowded and congested roads because of inefficient implementation of traffic rules. Motorists ignore the rules because they are not apprehended and can get away easily. This paper proposes an intelligent traffic system that is able to automatically detect and apprehend traffic violators, specifically motorists who either swerve or block the pedestrian lane. The system is designed by integrating three processes: violation detection, plate localization and plate recognition. The violation detection and plate localization were realized using genetic algorithm while the plate recognition process was performed using an artificial neural network. The recognition of the plate number is highly dependent on the position of the detected vehicle with respect to the camera. Thus, the recognized plate number will only be supplementary information about the violator; the physical attributes of the vehicle which is captured by the violation detection process will be the main information on the violator. Based on the results of 48 images tested, the overall system was able to detect the mentioned violations and to identify the plate number of the vehicles that were detected as traffic violators, with an average accuracy of 90.67%, and program runtime of 1.34 seconds.