Kinyua Wachira

Kinyua Wachira

Auto-generated profile, based on publicly available information

Kinyua Wachira

School of Engineering, University of Nairobi, P.O. Box 30197 Nairobi 00100, Kenya

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 November 2017
Verified
Conference Paper
Closed access
KWKinyua Wachira
Kinyua Wachira
Heywood Ouma AbsalomsHeywood Ouma Absaloms
Heywood Ouma Absaloms
JMJackson Mwangi Mbuthia
Jackson Mwangi Mbuthia
With rapid advances in the emerging and pervasive field of the Internet of Things, electrical and electronic engineering students need a robust grasp of embedded computational device principles to make a proper, appreciable and relevant contribution. This necessity is more pertinent in developing countries that are technologically lagging behind their developed counterparts. This paper analyses the teaching modalities and models used to teach embedded device theory. A new triumvirate (or three person group) hybrid blended learning model is proposed and used to instruct mid-level students in the Electrical and Information Engineering Department of the University of Nairobi; on two embedded device types: a custom board using an AVR Atmega88A series micro-controller and the Raspberry Pi single-board computer. Student circuit design ability and programming confidence prior and subsequent to the proposed method were used to assess the efficacy of the triumvirate model. From the preliminary iteration of the learning model program, it was found that the model is a viable method of ensuring persistent knowledge transfer of embedded design concepts and can be developed further.

1 September 2015
Verified
Conference Paper
Closed access
KWKinyua Wachira
Kinyua Wachira
This paper presents a novel scrutiny into the role of corrective terms in gradient-based demosaicking algorithms and how they can be used to improve image reconstruction. The terms analysed are the non-zero corrective term ε and a new inter-plane weighting term β. Two new techniques are proposed using these terms to highlight their contribution. Four performance metrics have been used (PSNR, CPSNR, SSIM and FSIM) for testing over four different image sets. Comparison of one the proposed algorithms is made with established current techniques and an improvement in image fidelity is observed over several image sets.

1 May 2015
Verified
Conference Paper
Closed access
KWKinyua Wachira
Kinyua Wachira
EMElijah Mwangi
Elijah Mwangi
Gradient-based Spatial demosaicking techniques have gained prominence in literature for their superior reconstruction capabilities. This paper presents a novel algorithm in this class with several key contributions. It introduces variables operating at various lattice levels of the Color Filter Array (CFA) data. It also employs a Square-On-Point (SoP) neighborhood, a corrective term and localized polling to reduce reconstruction errors thus improving image perception. The proposed algorithm is compared to other contemporary methods and an appreciable improvement in performance has been noted through Matlab simulation. To provide a robust analysis, two performance metrics (CPSNR and SSIM) are used over two distinct image sets.

1 September 2017
Verified
Conference Paper
Closed access
KWKinyua Wachira
Kinyua Wachira
Heywood Ouma AbsalomsHeywood Ouma Absaloms
Heywood Ouma Absaloms
The Raspberry Pi platform provides true equal opportunity-of-use in computing globally. From it, new affordable research and industrial opportunities have emerged. To fully realise its driving potential, particularly in developing countries, a persistent knowledge teaching process for the Raspberry Pi is necessary. This paper presents a quasi-flipped learning model (QFLM) intended to spark critical and creative thinking and foster an entrepreneurial spirit early in a student's tertiary education life-cycle. The model was applied as part of a computer science component for third year students in the Electrical & Information Engineering (E.I.E) Department of the University of Nairobi. Student programming confidence and completion of a time-bound project were used as model assessment metrics. It was found that the Raspberry Pi, coupled with the learning model, is a viable catalyst for local innovation and entrepreneurial opportunities.

1 September 2015
Verified
Conference Paper
Closed access
KWKinyua Wachira
Kinyua Wachira
EMElijah Mwangi
Elijah Mwangi
Gwanggil JeonGwanggil Jeon
Gwanggil Jeon
Spatial demosaicking methods are attractive due to their robust reconstruction capabilities and low computational load. This paper presents a new gradient-based spatial demosaicking algorithm with contributions from combinatorial geometry. It introduces the use of square tiling primitives called pentominoes to dictate the difference-term pairs combinations for gradient generation. The proposed algorithm is compared with current established methods. Reconstruction performance is objectively measured using three commonly documented metrics (MSE, SSIM and CPSNR). Three different image sets are used (Kodak, McMaster-IMAX and Condat) to evaluate demosaicking robustness. An significant improvement in performance over existing methods has been obtained through MATLAB simulation.

1 September 2017
Verified
Conference Paper
Closed access
KWKinyua Wachira
Kinyua Wachira
EMElijah Mwangi
Elijah Mwangi
GJGwanggil Jeon
Gwanggil Jeon
Recent studies have shown that panchromatic colour filter arrays possess better light intensity properties than the more prevalent Bayer array. This is an attractive property in integrated camera design. Demosaicking algorithms, however, focus primarily on Bayer-based cameras because of their simple demosaicking process. This paper presents a novel algorithm for use in the RGBW class of panchromatic arrays. It encodes light intensity information in a bayerisation conversion process and uses polyomino theory to define robust reconstruction interpolation weights. Performance of the proposed algorithm is assessed with established methods using two objective and commonly documented assessment measures (SSIM and CPSNR) over three (Kodak, USC-SIPI and Custom) image sets. An appreciable improvement is noted from MATLAB simulation, averaging 3dB above established techniques.