Teman Isyarat (Indonesian for "Sign Friend") — an AI-powered Indonesian Sign Language (BISINDO) recognition system. Built by a student team at Universitas Sebelas Maret (UNS) under the Hibah Jarprak program, in partnership with GERKATIN Solo.
| Repository | Purpose |
|---|---|
landmark extraction |
MediaPipe holistic landmark extraction pipeline — converts raw BISINDO video recordings into .npz landmark arrays (pose + hands) with ffmpeg-based data augmentation |
model pipeline |
GRU-based neural network with temporal attention — trains on extracted landmarks, evaluates with signer-independent cross-validation, exports TFLite models for mobile deployment |
android |
Flutter + Kotlin Android app — real-time BISINDO translation via CameraX and MediaPipe, running the TFLite model fully offline with temporal smoothing and skeleton overlay |
project manager |
Obsidian vault for project management — architecture decision records (ADRs), technical specifications, sprint tracking, and team documentation |
dataset |
Raw BISINDO video dataset collected with GERKATIN Solo |
frontend and backend |
Landing page and web platform |
The system recognizes 20 BISINDO vocabulary words (Central Java dialect):
| Category | Words |
|---|---|
| Pronouns | Aku, Kamu, Dia |
| Common | Salam, Terima Kasih, Maaf, Nama |
| Time | Hari Ini, Besok |
| Color | Merah, Kuning |
| Family | Ayah, Ibu |
| Count | Satu, Dua, Tiga |
| Other | Teman, Buku |
| Fruit | Apel, Pisang |
| Name | Role |
|---|---|
| Dunhill William | Project Lead |
| Fredy Ramadhan | R&D |
| Hany Wachidatul | R&D |
| Febrian Jamaludin | R&D |
| Ivan Wahyu | R&D |
| Kevin Marchelino | R&D |
| Mutia Rahman | R&D |
| Usrotun Saidah | R&D |
- GERKATIN Solo — Dataset collection, validation, and field testing
- Universitas Sebelas Maret (UNS) — Hibah Jarprak funding and academic support
- Built with MediaPipe, TensorFlow, and Flutter