Journal of ICT Research and Applications

Energy Consumption Prediction Using Data Reduction and Ensemble Learning Techniques

Marsa Thoriq Ahmada, Saiful Akbar — Tue, 27 Dec 2022 00:00:00 +0700

Building energy problems have various kinds of aspects, one of which is the difficulty of measuring energy efficiency. With current data development, energy efficiency measurements can be made by developing predictive models to estimate future building needs. However, with the massive amount of data, several problems arise regarding data quality and the lack of scalability in terms of computation memory and time in modeling. In this study, we used data reduction and ensemble learning techniques to overcome these problems. We used numerosity reduction, dimension reduction, and a LightGBM model based on boosting added with a bagging technique, which we compared with incremental learning. Our experimental results showed that the numerosity reduction and dimension reduction techniques could speed up the training process and model prediction without reducing the accuracy. Testing the ensemble learning model also revealed that bagging had the best performance in terms of RMSE and speed, with an RMSE of 262.304 and 1.67 times faster than the model with incremental learning.

Accuracy of Various Methods to Estimate Volume and Weight of Symmetrical and Non-Symmetrical Fruits using Computer Vision

Hurriyatul Fitriyah — Tue, 27 Dec 2022 00:00:00 +0700

Many researchers have used images to measure the volume and weight of fruits so that the measurement can be done remotely and non-contact. There are various methods for fruit volume estimation based on images, i.e., Basic Shape, Solid of Revolution, Conical Frustum, and Regression. The weight estimation generally uses Regression. This study analyzed the accuracy of these methods. Tests were done by taking images of symmetrical fruits (represented by tangerines) and non-symmetrical fruits (represented by strawberries). The images were processed using segmentation in saturation color space to get binary images. The Regression method used Diameter, Projection Area, and Perimeter as features that were extracted from the binary images. For symmetrical fruits, the best accuracy was obtained with the Linear Regression based on Diameter (LDD), which gave the highest R²(0.96 for volume and 0.93 for weight) and the lowest RMSE (5.7 mm³ for volume and 5.3 gram for volume). For non-symmetrical fruits, the highest accuracy for non-symmetric fruits was given by the Linear Regression based on Diameter (LRD) and Linear Regression based on Area (LRA) with an R² of 0.8 for volume and weight. The RMSE for LRD and LRA for strawberries was 3.3 mm³ for volume and 1.4 grams for weight.

Compact and Robust MFCC-based Space-Saving Audio Fingerprint Extraction for Efficient Music Identification on FM Broadcast Monitoring

Myo Thet Htun — Tue, 27 Dec 2022 00:00:00 +0700

The Myanmar music industry urgently needs an efficient broadcast monitoring system to solve copyright infringement issues and illegal benefit-sharing between artists and broadcasting stations. In this paper, a broadcast monitoring system is proposed for Myanmar FM radio stations by utilizing space-saving audio fingerprint extraction based on the Mel Frequency Cepstral Coefficient (MFCC). This study focused on reducing the memory requirement for fingerprint storage while preserving the robustness of the audio fingerprints to common distortions such as compression, noise addition, etc. In this system, a three-second audio clip is represented by a 2,712-bit fingerprint block. This significantly reduces the memory requirement when compared to Philips Robust Hashing (PRH), one of the dominant audio fingerprinting methods, where a three-second audio clip is represented by an 8,192-bit fingerprint block. The proposed system is easy to implement and achieves correct and speedy music identification even on noisy and distorted broadcast audio streams. In this research work, we deployed an audio fingerprint database of 7,094 songs and broadcast audio streams of four local FM channels in Myanmar to evaluate the performance of the proposed system. The experimental results showed that the system achieved reliable performance.

Translating SIBI (Sign System for Indonesian Gesture) Gesture-to-Text in Real-Time using a Mobile Device

Misael Jonathan, Erdefi Rakun — Sat, 31 Dec 2022 00:00:00 +0700

The SIBI gesture translation framework by Rakun was built using a series of machine learning technologies: MobileNetV2 for feature extraction, Conditional Random Field for finding the epenthesis movement frame, and Long Short-Term Memory for word classification. This high computational translation system was previously implemented on a personal computer system, which lacks portability and accessibility. This study implemented the system on a smartphone using an on-device inference method: the translation process is embedded into the smartphone to provide lower latency and zero data usage. The system was then improved using a parallel multi-inference method, which reduced the average translation time by 25%. The final mobile SIBI gesture-to-text translation system achieved a word accuracy of 90.560%, a sentence accuracy of 64%, and an average translation time of 20 seconds.

A Low Computational Cost RGB Color Image Encryption Scheme Process based on PWLCM Confusion, Z/nZ Diffusion and ECBC Avalanche Effect

Faiq Gmira, Wafae Sabbar, Said Hraoui — Sat, 31 Dec 2022 00:00:00 +0700

In this work, three sub-processes are serially integrated into just one process in order to construct a robust new image encryption scheme for all types of images, especially color images. This integration architecture aims to create a robust avalanche effect property while respecting the constraints of confusion and diffusion that have been identified by Claude Shannon as properties required of a secure encryption scheme. The performance of the proposed encryption scheme is measured and discussed with several analyses, including computational cost analysis, key space analysis, randomness metrics analysis, histogram analysis, adjacent pixel correlation, and entropy analysis. The experimental results demonstrated and validated the performance and robustness of the proposed scheme.

Cognitive Complexity Applied to Software Development: An Automated Procedure to Reduce the Comprehension Effort

Dinuka R. Wijendra, K. P. Hewagamage — Sat, 31 Dec 2022 00:00:00 +0700

The cognitive complexity of a software application determines the amount of human effort required to comprehend its internal logic, which results in a subjective measurement. The quantification process of the cognitive complexity as a metric is problematic since the factors representing the computation do not represent the exact human cognition. Therefore, the determination of cognitive complexity requires expansion beyond its quantification. The human comprehension effort related with a software application is associated with each phase of its development process. Correct requirements identification and accurate logical diagram generation prior to code implementation can lead to proper logical identification of software applications. Moreover, human comprehension is essential for software maintenance. Defect identification, correction and handling of code quality issues cannot be maintained without good comprehension. Therefore, cognitive complexity can be effectively applied to demonstrate human understandability inside the respective phases of requirements analysis, design, defect tracking, and code quality optimization. This study involved automation of the above-mentioned phases to reduce the manual human cognitive load and reduce cognitive complexity. It was found that the proposed system could enhance the average accuracy of requirements analysis and class diagram generation by 14.44% and 9.89% average accuracy incrementation through defect tracking and code quality issues compared to manual procedures.