IEEE

Robot-Mediated Explainability for CNN-Based Complex Emotion Recognition: Aligning Visual Attributions with Social Robot Behavior for Trust-Aware HRI

Abstract: In recent years, the development of socially interactive robots has heightened the need for interpretable and trustworthy models capable of recognizing complex human emotions in real time.
IEEE

Diffusion Model-Based Data Augmentation with Mel Spectrogram for Automatic Modulation Recognition in Underwater Acoustic Communication

Abstract: Automatic modulation recognition (AMR) in underwater acoustic (UWA) communications is limited by scarce training data and channel impairments such as multipath propagation, Doppler shifts, ...
GitHub

SynthToSoul (Human vs. AI Music Detector and Top-K Similar Songs Recommender)

Overview SynthToSoul is a full-stack audio analysis application designed to distinguish between Human-Made and AI-Generated music. By combining a custom Deep Learning classifier with audio ...
IEEE

Human Daily Activity Recognition using Spectrogram and 2D-CNN Through Doppler Radar

Abstract: The identification and analysis of human daily activities has garnered substantial attention in recent years, driven by its expansive applications in areas including healthcare, surveillance ...
IEEE

TifinNet: CNN–Transformer Hybrid Architecture for Tifinagh Handwritten Recognition

Abstract: Handwritten Optical Character Recognition (OCR) for the Tifinagh script remains a challenging task due to the script’s geometric complexity, high similarity among rotated characters, and the ...
IEEE

Detecting Deep Fake Audio: A Hybrid CNN-LSTM Approach Leveraging Spectrogram Analysis

Abstract: The rapid progress in audio synthesis technologies allows the generating of artificial audio deep-fakes that sound convincingly real, thus causing security, privacy, and authenticity ...
IEEE

Lightweight and Personalized Single-Eye Emotion Recognition via CNN-SNN Spatiotemporal Learning and Memory-Inferred Event Features

Abstract: Emotion recognition is essential for improving user experience and interaction quality in human-centered applications. While recent studies have leveraged both event and traditional cameras ...
IEEE

Mel Spectrogram Based CNN with Attention Mechanisms for Robust Music Genre Classification

Abstract: Music genre classification is an important task in music information retrieval with uses in recommendation, indexing and streaming. This paper proposes a CNN-based framework that classifies ...
IEEE

CKADisor: Central Kernel-Aligned Distillation for Efficient Event-Driven Object Recognition

Abstract: Spiking neural networks (SNNs) are one of the best practices for efficient event-driven object recognition. To achieve high recognition accuracy, existing methods generally accumulate ...
IEEE

Weak-Mamba-UNet: Visual Mamba Makes CNN and ViT Work Better for Scribble-based Medical Image Segmentation

Abstract: Medical image segmentation is increasingly reliant on deep learning techniques, yet the promising performance often come with high annotation costs. This paper introduces Weak-Mamba-UNet, an ...
IEEE

Deep Spectrogram-Based EEG Decoding of Visual Stimuli Using a Novel CNN Architecture

Abstract: Electroencephalography EEG - based decoding of visual stimuli has gained traction in computational neuroscience and brain-computer interface applications. This study explores the feasibility ...
IEEE

A Progressive Fusion Framework of Rotationally-Enhanced 3D CNN and GCN-DTW for Table Tennis Action Recognition and Quality Evaluation

Abstract: Most existing studies on table tennis training focus on either action recognition or scoring alone, lacking a systematic way to model both tasks together, which limits the ability to provide ...