In this paper, an image binarization method for separating text from the background of degraded textual images is proposed. This proposed methods are based on combination of Window Tracking Method (WTM) and Dynamic Image Histogram (DIH). The WTM and DIH methods work on an image that has been pre-processed. The WTM method searches for the largest pixel value in a 3 × 3 window up to a maximum of five tracking steps, while the method searches for a definite frequency between the two highest values in the image histogram. We test proposed method on DIBCO dataset and self-collection faded manuscripts. The experimental results show that our proposed method outperforms state of the art methods.
Oleh sebab kejadian jenayah bersenjata api semakin berleluasa, pengecaman senjata api yang digunakan oleh penjenayah amat diperlukan sebagai bahan bukti dalam mahkamah. Beberapa sistem pengecaman senjata api telah diutarakan sebagai pengganti kepada cara penyiasatan tradisional yang amat bergantung kepada kepakaran ahli balistik. Pemetakan rantau tumpuan (ROI) berdasarkan kedudukan titik sauh (PAP) sempadan bulatan kesan pin peletup pada tapak kelongsong peluru merupakan langkah yang amat penting dalam sistem pengecaman senjata api automatik. Walau bagaimanapun, kaedah yang digunakan dalam kajian lepas bagi mengesan PAP sempadan bulatan tersebut adalah sangat kompleks dan memerlukan masa pemprosesan yang panjang. Kajian ini menerokai algoritma yang efisien dan berkemampuan untuk mengesan PAP sempadan bulatan secara automatik. Algoritma yang diutarakan merupakan gabungan daripada penapis penajaman reruang, penormalan histogram, pengambangan dan penganggar kuasa dua terkecil tak berpemberat. Dua kaedah pengambangan yang terkenal telah diuji dan dibandingkan, iaitu kaedah pengambangan berasaskan pengelompokan dan kaedah berasaskan entropi. Di samping itu, penerokaan kesan saiz dan bentuk ROI terhadap kadar pengelasan senjata api turut dipersembahkan. Sebanyak 747 imej kesan pin peletup jenis sempadan bulatan peletup yang dihasilkan oleh lima pucuk pistol yang berlainan daripada jenis yang sama digunakan untuk menguji algoritma yang diutarakan. Kadar pengelasan imej kesan pin peletup yang memberangsangkan (> 95%) telah dicapai dengan algoritma yang dicadangkan. Kajian juga mendapati bahawa saiz dan bentuk pemetakan ROI mempunyai kesan langsung terhadap kadar pengelasan senjata api.
The Green Zone of Emergency Department Hospital Universiti Sains Malaysia (EDHUSM) which provides treatment for
non-critical cases contributes partly to the hustle and bustle in the emergency department. The imbalance of doctors
and nurses with the patient ratio which forms the resources’ bottleneck further results to the long patients’ waiting time
especially after the office hours and during weekends and public holidays. Collectively, this disproportion and bottlenecks
roots up the current problem faced by Green Zone EDHUSM which constantly fails to achieve the KPIs set by the hospital.
Henceforth, this study focuses on the best resource allocation of doctors and nurses for shifts during the weekdays and
for shifts during weekends and public holidays. The hybrid method of Discrete Event Simulation, and Data Envelopment
Analysis models such as BCC-input oriented and Super-Efficiency, were deployed to obtain the best resource allocation
for the two groups of shift. The method produced a series of resources allocation alternatives for doctors and nurses
with a total of 64 alternatives for weekdays and 729 alternatives for weekends and public holidays. The results show that
the best allocation for doctors and nurses during weekdays are three doctors and three nurses serving for every shift,
while during weekends and public holidays, a combination of four doctors and four nurses for every shift are the best.
The proposed combinations have reduced the average waiting time, optimized the utilization of doctors and nurses, and
managed to increase the number of patients served during weekdays, weekends and public holidays.