Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userComparative Study between background subtraction and Gaussian Mixture Model algorithms for outdoors videos
دراسة مقارنة بين طريقتي طرح الخلفية و نموذج مزيج غاوص المستخدمتين للتخلص من الخلفية في فيديوهات ملتقطة في الهواء الطلق
1057
1 40
5.0
(
1
)
Created by
Shamra Editor
Ask ChatGPT about the research
يعد التخلص من الخلفية الخطوة الأولى في أغلب منظومات رؤية الآلة الرامية إلى عزل كائن معين و التعرف عليه و استخلاص المعلومات منه خصوصاً تلك التي تعتمد على تتبع الحركة (كالتعرف على الإنسان أو تتبع السيارات ...الخ). و تشكل الفيديوهات الملتقطة في الهواء الطلق تحدياً أمام التخلص من الخلفية نظراً لحدوث تغيرات غير مرغوب فيها كتأثير الرياح و الإضاءة و الظروف المناخية و غيرها. يقدم هذا البحث دراسة مقارنة بين أبسط طريقة للتخلص من الخلفية (طريقة طرح الخلفية) و طريقة نموذج مزيج غاوص التي تعتبر أكثر تلك الطرق شيوعاً. تتمهذه المقارنة على أساس قدرة كل طريقة على التخلص من الخلفية في فيديوهات ملتقطة في الهواء الطلق و بشكل خاص في حالة وجود و عدم وجود ظل بالإضافة لتحديات أخرى كحركة الأجسام في الخلفية و تأثير الرياح و عدم استقرار الكاميرا. و الانطلاق من نتائج تلك المقارنة لاختيار الطريقة المناسبة لتتبع الحركة في ظروف و شروط مختلفة.
Background subtraction (BS) is the first step of various computer vision application specially those depending on motion tracking such as (car tacking, human recognition…etc.).Indeed, videos captured outdoors may contain a lot of undesirable changes ‘wind impact, illumination changes, weather conditions and others ’, generate numerous false positives. This paper presents comparison between the simplest method for background extraction (background subtraction) and Gaussian Mixture Model which is common method in outdoors videos. These two method are then compared based on the ability of each one to detect moving object in outdoors videos especially with presence and absence of shadow in addition to other challenges like object movement in background, wind effect and camera instability. The results of this comparison is used to determine the suitable method for each state.
References used
B. a. S. B. Horn, "Determining optical flow.," Artificial Intelligence, vol. 17, no. 1, pp. 185- 203, 1981
Q. Z. a. J. Aggarwal, "Tracking and classifying moving objects from video," in Performance Evaluation of Tracking Systems Workshop, 2001
B. A. Smith, Determination of Normal or Abnormal Gait Using a Two Dimensional Video Camera, Blacksburg,Virginia: Polytechnic Institute and State University, 2007
rate research
Read More
Dental caries increasingly affects the children nowadays. Statistics
concerning caries prevalence aim to guide and improve the
prevention of dental caries.
Objective: The purpose of this study is to assess the prevalence of
dental caries in the posterior primary teeth depending on the
presence of decay, missing due to caries or filling.
This paper, intended for the ISA-17 Quantification Annotation track, provides background information for the shared quantification annotation task at the ISA-17 workshop, a.k.a. the Quantification Challenge. In particular, the role of the abstract an
d concrete syntax of the QuantML markup language are explained, and the semantic interpretation of QuantML annotations in relation to the ISO principles of semantic annotation. Additionally, the choice is motivated of the test suite of the Quantification Challenge, along with the suggested markables for the sentences of the suite.
The pelvis anatomy and position, interact with the lumbar spinal organization in shape and position to regulate the sagittal balance between both the lumbar spine and pelvis. Sagittal lumbar-pelvic balance of the human body may be analyze by a pelvic
parameters ( pelvic incidence (PI), and sacral slope (SS)) and shape of the spine (lumbar lordosis (LL) and spinal morphologic type). Hamstring muscle tightness (HMT) causes rotating backward the pelvis (retroversion) which decrease lumbar lordosis. We measure PI, SS, lumbar lordosis and spine type on lateral radiographs of 21 patients with lumbar disc herniation (LDH) and 50 spinal healthy subjects and evaluate HMT by straight-leg-raising test (SLR-test). The lumbar-pelvis parameters for patients with DH demonstrated to have a mean PI, SS and LL equal to 48.85°, 37.9°and 47.28°, respectively, versus 52.24°, 42.88° and 59.24°, respectively, for the control group. Patients with a LDH were characterized to have PI, SS and LL significantly lower than the control group. Also DH group has a mean SLR-test equal to 74.85°, significantly lower than control group (82.24°). We notice also that spine type I and type II are more frequently in DH group. All results give that DH group has vertical spine more than control one, which mean that the DH patients have vertical pressure on lumbar discs more than normal subjects.
Cross-attention is an important component of neural machine translation (NMT), which is always realized by dot-product attention in previous methods. However, dot-product attention only considers the pair-wise correlation between words, resulting in
dispersion when dealing with long sentences and neglect of source neighboring relationships. Inspired by linguistics, the above issues are caused by ignoring a type of cross-attention, called concentrated attention, which focuses on several central words and then spreads around them. In this work, we apply Gaussian Mixture Model (GMM) to model the concentrated attention in cross-attention. Experiments and analyses we conducted on three datasets show that the proposed method outperforms the baseline and has significant improvement on alignment quality, N-gram accuracy, and long sentence translation.
The study aims at comparing ARIMA models and the exponential
smoothing method in forecasting. This study also highlights the special
and basic concepts of ARIMA model and the exponential smoothing
method.
The comparison focuses on the ability
of both methods to forecast
the time series with a narrow range of one point to another and the time
series with a long range of one point to another, and also on the different
lengths of the forecasting periods. Currency exchange rates of Shekel to
American dollar were used to make this comparison in the period
between 25/1/2010 to 22/10/2016. In addition, weekly gold prices were
considered in the period between 10/1/2010 to 23/10/2016. RMSE
standard was used in order to compare between both methods. In this
study, the researcher came up with the conclusion that ARIMA models
give a better forecasting for the time series with a long range of one point
to another and for long term forecasting, but cannot produce a better
forecasting for time series with a narrow range of one point to another as
in currency exchange prices.
On the contrary, exponential smoothing method can give better
forecasting for Exchange Rates that has a narrow range of one point to
another for its time series, while it cannot give better forecasting for long
term forecasting periods
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
