My research interests are (but not limited to) Computer vision, Pattern recognition, UI/UX and Animal re-ID. I also doodles.

Interested in pursuing an Msc or a PhD? Drop me an email at mhihipni at unimas dot my

---

Projects:

Design of A Deep Learning Model with Attention Mechanism for Biometric Re-identification of Green Sea Turtles in Long-term Tracking Scenario

We work with Sarawak Forestry Corp. to develop an automated biometric re-identification prototype for nesting green sea turtles. Instead of using plastic/metal tags to re-identify individuals, our solution utilises a motion-activated, downright-facing camera to capture images of nesting female sea turtles. We trained deep learning models to extract the ROIs and infer individuals' ID. We have published three articles: i. IJAI2024, in International Journal of Artificial Intelligence, 13(2), pp. 2354-2363, ii. IJCDS2024, in International Journal of Computing and Digital Systems, 16(1), pp. 201-211, and iii. SCI.DATA2024, in Nature's Scientific Data, 11, 1441 (2024).

Who Danced Better? Ranked Dance Video Dataset and Pairwise Skill Annotation Method

We developed a pairwise skill annotation method for TikTok dance challenges. The method encodes the subject's consecutive 2D poses into string values to preserve the order and tempo (of micro movements). The article is published: IJAIN2023, in International Journal of Advances in Intelligent Informatics 9(1), pp. 96-107, 2023. Dataset is available here: link. A 2nd article utilising the same dataset has also been published: IJ-AI2023, in International Journal of Artificial Intelligence 12(2), pp. 602-609, 2023.

A Novel Keypoint-based Recognition Pipeline for Iban Mat Motifs utilising Adaptive Smoothing, a PhD student project

The first article, i.e., IJEECS2021, provides evidence that we can use local feature descriptors to distinguish between classic Iban mat motif classes. Silvia's 2nd article, i.e., IJ-AI2023, introduces a classification framework that utilise adaptive smoothing to exclude small and non-discriminative features from affecting the classification results.

AR-PHOTOBOOTH© available for purchase)

More details,

MO-LEARN©: Kinaesthetic Learning (available for purchase)

More details,

MO-WALL©: Interactive Kiosk with Gesture Recognition (available for purchase)

More details

Biometric identification of sea turtles (Chelonia mydas)

Hipiny, I., Ujir, H., Mujahid, A., and Yahya, N.K. (2018), Towards Automated Biometric Identification of Sea Turtles (Chelonia mydas), Journal of ICT Research and Applications, 12(3), 256-266, doi:10.5614/itbj.ict.res.appl.2018.12.3.4, Link

Activity recognition using gaze

I worked on modeling activities as a record of fixated visual landmarks, i.e., Gaze regions (See video), [BRISTECHREPO12]. The Gaze regions also prime locations of natural temporal cuts inside an egocentric video, [CITA15], and their properties are indicative of the subject's task performance, [ICSIPA17a]. My PhD (2014) thesis is available here.

---

Current Postgraduates

Name, Lvl	Thesis title	Remarks
Soria S., MSc	Automatic segmentation of Region-of-Interest (ROI) in nesting Green sea turtle videos	Feb 2022 - Now, [IJCDS2024]
Leghari I., PhD	Artificial Intelligence Approaches for Identification and Evaluation of Intellectual Disabled Down Syndrome People of Pakistan: A Model Based Study	Oct 2022 - Now, Co-sup w/ AP Dr Hamimah
Ghazvini A., PhD	Preserving Visual Integrity in Content-Aware Image Retargeting	Apr 2024 - Now

---

Graduated

Name, Lvl	Thesis title	Remarks
Peter M., MSc	Facial Expression Synthesis using Kernel Approach	Co-sup w/ Dr Jacey-Lynn Minoi
Joseph S., PhD	Iban’s Plaited Mat Motifs Recognition using Invariant Image Features	Thesis, Now at MOH Malaysia
Zakry K.A., MSc	Deep Learning Model with Attention Mechanism for Biometric re-ID of Green Sea Turtles in Long-term Tracking Scenario	Thesis

---

Datasets

My Kaggle datasets link.

---

Miscellaneous

1. Freeglut - I use this instead of GLUT in my Computer Graphics’ class
2. MinGW
3. NeHe - The best resource site for OpenGL
4. Colah’s excellent LSTM tutorial
5. An excellent Youtube video on JPEG compression, DCT ***