In Bangladesh, there is an alarming shortage of mental health professionals, with only 1.17 per 100,000 population and just six child psychiatrists nationwide. Moreover, there is no culturally adapted Bengali-language tool for early screening of abuse-related psychological trauma in children.
We present ShishuRaksha AI, a decision-support (not diagnostic) framework that fuses four screening modalities: validated questionnaires (SDQ, CPSS), Bengali narrative text, House-Tree-Person (HTP) drawing features, and facial affect analysis. The fusion is training-free, clinically weighted, utilizes cross-modal attention, and includes a single-modality override rule.
Each risk score is explained through clinically weighted, perturbation-based additive attribution and rendered as a bilingual (Bangla/English) report with referral routing to national child-protection services (OCC, DSS, NMHH) under the Children Act 2013.
Due to ethical constraints, we cannot collect clinical datasets of abused children, so we introduce a noise-aware synthetic benchmark (500 cases, 116 positive [23.2%], four deliberate noise layers, literature-grounded HTP priors) and evaluate tree-ensemble surrogates of the fusion design (excluding the facial channel) under 5-fold stratified cross-validation.
The fused model achieves an AUC of 0.874 [0.834-0.908], compared to 0.756 [0.705-0.803] for an SDQ-only baseline, with ablation, operating-point, subgroup, and calibration analyses. We openly state all limitations, including synthetic-only data, no held-out set, text-feature circularity, and an urban-rural subgroup gap.
This work is a feasibility study and a design contribution toward ethically deployable child-protection screening in low-resource settings.
Blogger's Review: This study presents an innovative solution for addressing child abuse trauma screening in resource-limited settings, leveraging multimodal data fusion to enhance screening capabilities. Despite facing ethical and data limitations, the methodological design lays a solid foundation for future applications, making it a noteworthy contribution to the field.