Personalized Fall Detection by Balancing Data with Selective Feedback Using Contrastive Learning

Awatif Yasmin; Tarek Mahmud; Sana Alamgeer; Anne H. H. Ngu

cs.LGMar 17, 2026

Personalized Fall Detection by Balancing Data with Selective Feedback Using Contrastive Learning

Authors: Awatif Yasmin, Tarek Mahmud, Sana Alamgeer, Anne H. H. Ngu

Organizations: Texas State University · Texas A&M University - Kingsville

Abstract

Personalized fall detection models can significantly improve accuracy by adapting to individual motion patterns, yet their effectiveness is often limited by the scarcity of real-world fall data and the dominance of non-fall feedback samples. This imbalance biases the model toward routine activities and weakens its sensitivity to true fall events. To address this challenge, we propose a personalization framework that combines semi-supervised clustering with contrastive learning to identify and balance the most informative user feedback samples. The framework is evaluated under three retraining strategies, including Training from Scratch (TFS), Transfer Learning (TL), and Few-Shot Learning (FSL), to assess adaptability across learning paradigms. Real-time experiments with ten participants show that the TFS approach achieves the highest performance, with up to a 25% improvement over the baseline, while FSL achieves the second-highest performance with a 7% improvement, demonstrating the effectiveness of selective personalization for real-world deployment.

Figures & tables

Figure 3: (a) Overview of the pipeline for training personalized models (

M_{Pi}

) & (b) SNN architecture

Figure 5: Precision and recall over time for models trained with selective data vs. all feedback data

Participant ( $i$ )	Baseline	Personalized Models ( $M_{Pi}$ )
Participant ( $i$ )	$M_{O}$	$M_{TFS}$	$M_{TL}$	$M_{FS}$
Participant 1	0.69	0.89	0.72	0.79
Participant 2	0.71	0.9	0.69	0.78
Participant 3	0.76	0.92	0.67	0.8
Participant 4	0.68	0.93	0.7	0.76
Participant 5	0.73	0.94	0.75	0.74

Table 1: Real-time evaluation results of 10 participants using all 3 approaches of retraining vs

M_{O}

(bold values indicate the highest F1 score achieved across models for a given participant)

Dataset	Accuracy
SmartFallMM	91%
HAR	94%
CZU-MHAD	85%
K-Fall	86%

Table 2: Accuracy of SNN on Different Datasets

Figure 6: Clustering accuracy for each activity

Input for SNN	Precision	Recall	F1-Score
Similarity metrics	0.92	0.9	0.91
Basic statistical features (min, max, mean, std)	0.75	0.55	0.63
Raw time-series data	0.63	0.72	0.67

Table 3: Comparing the performance of SNN with different inputs

Training Data	F1-Score
Initial dataset + 5% feedback data	0.80
Initial dataset + 10% feedback data	0.85
Initial dataset + 15% feedback data	0.88
Initial dataset + 20% feedback data	0.91
Initial dataset + 25% feedback data	0.89

Table 4: Model performance across different amounts of feedback data while personalizing

Fall Detection

Fall detection identifies falls and impact events from video, body-worn sensors, radar, and other signals to support timely assistance.

See all 20 papers

Class-Aware Adaptive Differential Privacy in Deep Learning for Sensor-Based Fall Detection
May 3, 2026Joydeb Kumar SanaDifferentially Private Stochastic Gradient Descent Privacy-Preserving ML
Institute of Information and Communication Technology, Bangladesh University of Engineering and Technology, Dhaka, Bangladesh
Beyond Simulated Benchmarks: Evaluating Motion Representations for Fall Detection Under Real-World Data Scarcity
Aug 13, 2026Timilehin B. Aderinola, Ilaria D'Ascanio, Luca Palmerini +5Fall Detection Representation Learning
School of Computer Science, University College Dublin (UCD), Ireland · Department of Electrical, Electronic and Information Engineering “Guglielmo Marconi”, University of Bologna, Italy · Department of Clinical Gerontology, Robert Bosch Hospital, Stuttgart, Germany +2
FALLNEST: Dual-Domain Dataset Condensation of Imbalanced IMU Data for Wearable Fall Detection
Oct 13, 2026Rashi Niyas P, Ritika Jain, Hitika TiwariFall Detection Dataset Condensation
Indian Institute of Technology Madras · Indian Institute of Technology Guwahati

May 3, 2026J/K move · Enter open · S save

Class-Aware Adaptive Differential Privacy in Deep Learning for Sensor-Based Fall Detection

Joydeb Kumar Sana

Institute of Information and Communication Technology, Bangladesh University of Engineering and Technology, Dhaka, Bangladesh

Differentially Private Stochastic Gradient Descent Privacy-Preserving ML Differential Privacy

Open paper AskPDF arXiv

Fall detection is a critical task in healthcare, particularly for elderly people. Timely fall detection and treatment can prevent severe injuries. Sensor-based activity data can be used to detect fall. However, this data are highly sensitive and raises significant privacy concerns. Existing privacy approaches apply uniform noise across all training samples, which affects the prediction performance. To address this limitation, we propose a Class-Aware Adaptive Differential Privacy (CA-ADP) framework integrated with a hybrid 3D Convolutional Neural Network and Bidirectional Long Short-Term Memory (3D CNN-BiLSTM) architecture. The CA-ADP mechanism dynamically adjusts the magnitude of noise added to gradients based on the class composition of each mini-batch. This process ensures privacy while mitigates performance degradation. We formally analyze the

(ε,δ)

-Differential Privacy guarantee and provide a privacy-utility trade-off analysis. The proposed method is evaluated on three public benchmark datasets, namely SisFall, UP-Fall, and MobiAct. The experimental results show that the proposed privacy model achieves improvements of 3.3%, 8.5%, and 7.5% over the conventional privacy-based model in terms of F-score for the SisFall, UP-Fall, and MobiAct datasets, respectively. Comparisons with prior studies show that the CA-AD based framework achieves competitive performance and provides formal privacy guarantees, which are largely overlooked in existing studies. Wilcoxon signed-rank tests confirm that the proposed mechanism consistently outperforms conventional differential privacy. Those results establish the proposed CA-ADP framework as an effective approach to privacy-preserving fall detection in real-world healthcare settings.

Abstract

Fall Detection

Explore similar work

Class-Aware Adaptive Differential Privacy in Deep Learning for Sensor-Based Fall Detection

Beyond Simulated Benchmarks: Evaluating Motion Representations for Fall Detection Under Real-World Data Scarcity

FALLNEST: Dual-Domain Dataset Condensation of Imbalanced IMU Data for Wearable Fall Detection