Showing papers on "Dark-frame subtraction published in 2004"

Speech enhancement by noise masking

Abstract: The invention provides a method and apparatus for noise reduction of a signal, for example speech enhancement of a speech signal. The method involves a two-stage algorithm comprising performing a preprocessing first spectral subtraction to remove tonal noise and generate a tonal noise removed signal, and performing a second spectral subtraction to remove noise from the said tonal noise removed signal. In both spectral subtraction stages noise is not removed completely but only to a level below an audible threshold in order to avoid unwanted artifacts.

Adaptive hybrid mean and median filtering of high-ISO long-exposure sensor noise for digital photography

Abstract: This paper presents a new methodology for the reduction of sensor noise from images acquired using digital cameras at high- International Organization for Standardization (ISO) and long- exposure settings. The problem lies in the fact that the algorithm must deal with hardware-related noise that affects certain color channels more than others and is thus nonuniform over all color channels. A new adaptive center-weighted hybrid mean and median filter is formulated and used within a novel optimal-size windowing framework to reduce the effects of two types of sensor noise, namely blue-channel noise and JPEG blocking artifacts, common in high-ISO digital camera images. A third type of digital camera noise that affects long-exposure images and causes a type of sensor noise commonly known as ''stuck-pixel'' noise is dealt with by pre- processing the image with a new stuck-pixel prefilter formulation. Experimental results are presented with an analysis of the perfor- mance of the various filters in comparison with other standard noise reduction filters. © 2004 SPIE and IS&T. (DOI: 10.1117/1.1668279)

Authenticating corrupted face image based on noise model

Abstract: In this paper, we propose a method for authenticating corrupted face images based on noise model. The proposed method first generates corrupteed images by controlling nois parameters in the training phase. The corrupted images and noise parameters are represented by a linear combination of prototypes of the corrupted images and the noise parameters. With the corrupted image and an original image, we can estimate noise parameters of the corrupted face image in the testing phase. Then, we can make a synthesized face image from the original face image with the estimated noise parameters and verify it with the corrupted face image. Our experimental results show that the proposed method can estimate noise parameters accurately and improve the performance of face authentication.

Detection of moving shadows using mean shift clustering and a significance test

Abstract: An algorithm that discriminates moving objects from their shadows is presented. Starting from the change mask of an image sequence, first of all the changed area is divided into subregions consisting of pixels with similar colour properties. This is done using the mean shift algorithm, which is very powerful in non-parametric clustering of data. In a second step a significance test is performed to classify each image pixel inside the change mask into one of the classes foreground or shadow. To do this a straightforward image model is used where the grey level of a foreground pixel covered by a shadow is given by the product of the corresponding background pixels' grey-level and a constant value. Assuming that fore- and background images are corrupted by Gaussian white noise, a significance test is derived which classifies all pixels inside the change mask. In the third step global and local information from the first and second steps are combined. For each region inside the change mask it is examined if the majority of pixels survived the second step. If this is the case, the whole region is kept for the final moving object mask, if not the region is set to zero.

Video comfort noise addition technique

Abstract: A decoding arrangement for decoding pictures in an incoming video stream includes a noise generator for adding a dither signal containing random noise to the pictures after video decoding, to improve the subjective video quality. The noise generator adds noise to each pixel in an amount correlated to additive noise of pixels in a prior picture, either a previously displayed picture (i.e., a previously decoded picture to which noise has been added), or a previously decoded picture.

Analysis and processing of CCD noise

Abstract: In order to improve CCD performance, the noise composition in CCD image is analyzed according to the operating principle of CCD, and CCD noise is categorized Correspending measures are introduced to deal with different noise Signal processing schemes of double-correlated sampling,dual slope integration and clamp sample are given to eliminate readout noise in this paper As a result, the SNR of CCD is improved

Radiometric calibration of CCD sensors: dark current and fixed pattern noise estimation

Abstract: The irradiance measurement performed by vision cameras is not noise-free due to both processing errors during CCD fabrication and the behaviour of the electronic device itself. A proper characterization of sensor performance, however, allows either removing the resulting noise from the image or accounting for it within image processing algorithms. This paper proposes new methods for estimating the distribution parameters of two of such sources of noise: dark current and the so-called fixed pattern noise. Since both methods require knowledge about the scene illumination, an estimation method using a calibrating sphere is also presented. This method models illumination as the combination of directional and ambient lighting. Experimental results can be found at the end of the paper.

Low noise sample and hold circuit for image sensors

Abstract: An image sensor includes a plurality of pixels for capturing incident light that is converted to a signal representing an image; wherein noise is combined with a signal representing both the image and a reset level; a plurality of dark reference pixels that generate noise that substantially correspond or equally correspond to the noise in the image and reset level; and a sample and hold circuit that reads out the image signal and cancels or substantially cancels out the noise from the image signal and reset level by canceling the noise from image and reset level with the noise generated from the dark reference pixels.

Noise suppression processing method, apparatus and program

Abstract: A band-limited image signal generating unit decomposes an input image signal into a plurality of band-limited image signals, each representing an image having different frequencies. An index value obtaining unit obtains an index value indicating a level of noise suppression based on data representing spatial frequencies as well as an evaluation value representing local contrast at a pixel of interest in band-limited images and data representing an X-ray dose. A noise suppression processing unit performs noise suppression processing on each pixel of each of band-limited images based on the index value. A processed image generating unit obtains a processed image, in which noise has been suppressed, by adding signals obtained by subtracting processed band-limited image signals, of which noise has been suppressed, from the band-limited image signals together, extracting a noise image signal, and subtracting the noise image signal from the input image signal.

Image sensor with dark signal reduction

Abstract: An image sensor that has a temperature sensor. The temperature sensor senses the temperature of a pixel array of the image sensor. The sensed temperature is used to scale a dark frame image generated by the pixel array. The scaled dark frame image is subtracted from a light image frame generated by the pixel array. The scaled dark image frame compensates for temperature variations in the pixel array. The scaled dark image frame may be generated by multiplying the dark frame by a scale factor(s). The scale factor may be computed from an equation or determined from a look-up table. The equation or look-up table may compensate for thermal gradients across the pixel.

Locally adaptive nonlinear noise reduction

Abstract: An imaging scanner (10) acquires imaging data. A reconstruction processor (30) reconstructs the imaging data into an unfiltered reconstructed image. A local noise mapping processor (64, 120, 136, 140, 142, 152) generates a noise map (68, 68', 68'') representative of spatially varying noise characteristics in the unfiltered reconstructed image. A locally adaptive non linear noise filter (60) differently filters different regions of the unfiltered reconstructed image in accordance with the noise map (68, 68', 68'') to produce a filtered reconstructed image.

Automatic assessment and reduction of noise using edge pattern analysis in non-linear image enhancement

Abstract: Noise is the primary visibility limit in the process of non-linear image enhancement, and is no longer a statistically stable additive noise in the post-enhancement image. Therefore novel approaches are needed to both assess and reduce spatially variable noise at this stage in overall image processing. Here we will examine the use of edge pattern analysis both for automatic assessment of spatially variable noise and as a foundation for new noise reduction methods.

An electronic calibration scheme for logarithmic CMOS pixels

Abstract: Logarithmic cameras have the wide dynamic range required to image natural scenes and encode the important contrast information within the scene. However, the images from these cameras are severely degraded by fixed pattern noise. Previous attempts to improve the quality of images from these cameras by removing additive fixed pattern noise have led to disappointing results. Using a three parameter model for the response of logarithmic pixels, it is concluded that the residual fixed pattern noise in these images is caused by gain variations between pixels. In order to reduce the effects of these variations a new type of readout circuit has been designed. However, even with this readout circuit high quality images will only be obtained if each image is corrected to remove the effects of both gain and offset variations. Measurement results are presented that show that the quality of the output from logarithmic pixels is significantly improved if a procedure is used which corrects for both types of variations. In fact with this procedure the contrast sensitivity of the logarithmic pixels becomes comparable to that of the eye over five decades of input illumination intensity.

Noise removal method, storage medium having stored therein noise removal processing program and noise removing apparatus

Abstract: Image data are obtained by photographing a target image containing fixed pattern noise and a noise image photographed in a dark state and having recorded therein position information indicating the position of the fixed pattern noise are obtained. The position of each pixel indicating a color information value equal to or greater than a threshold value is extracted from the noise image and the color information at the pixel at the corresponding position in the target image is substituted with a median value of color information values at surrounding pixels so as to eliminate the fixed pattern noise from the target image.

Digital image restoration by exposure-splitting and registration

Abstract: With the explosion of digital imaging systems, high image noise levels, particularly motion blur and sensor noise, limit the applications of mobile cameras. In this paper, a novel restoration approach is developed. By uniformly splitting the exposure time and averaging the multiple under-exposed captures, a single image is obtained in which both types of noise are mitigated. Comprehensive experiments demonstrate the successful image restoration under various exposure conditions. Exposure-splitting is further optimized to fully recover the image quality under severe motion blur. This technique can be easily implemented into contemporary digital imaging systems.

Noise removing device for image sensor

Abstract: The noise removing device for an image sensor according to the present invention is for reducing the noise in the output of the image sensor. The device comprises: a noise generator for generating the noise in common-mode with respect to the output of the image sensor; and a differential amplifier which amplifies the difference between the output of the image sensor and the output of the noise generator. By canceling the common-mode signals, 1/f noise, low-frequency noise, high-frequency noise, shot noise, beat noise and the like, which are asynchronous with the horizontal synchronizing signals, can also be reduced.

Apparatus and method of measuring noise in a video signal

Abstract: An apparatus and method of measuring noise in video signals includes a high-frequency component determination part that detects a high-frequency component value of a first image to measure noise in a blockwise unit, a spatial filter that filters the first image in the blockwise unit, by applying different filtering methods according to the high frequency component value, and outputting the filtered image as a second image, a motion compensation error determination part that determines a presence of a motion compensation error by comparing a first difference between corresponding pixel values of the first image and the second image with a second difference between corresponding pixel values of the first image and a third image which is motion-compensated image derived from the first image, and a noise calculator that measures noise with reference to the second difference between the corresponding pixel values of the first image and the third image for pixels determined by the motion compensation error determination part to have no motion compensation error.

Ordering Threshold Switching Median Filter

Abstract: This paper presents an ordering threshold switching median filter to solve the contradiction between noise attenuation and image detail preserving. From the ordering information of the pixels in the window, and based on extremum median filtering the image corrupted by impulse noise is divided into three pixel classes, that is, noise pixels, edges and details, and smooth regions. With the statistic of a lot of standard images tested, the parameters of the classifier are properly chosen in order to deal with most images adaptively. Then switching median filtering is applied with the classifier. Therefore the smooth regions and noise pixels are filtered by median filters that have a good noise removing capability, especially with the 'salt and pepper' noise. However, most of the edges and details of the image are untouched, so that the restored image can keep details even in variable magnitude impulse noise conditions. A comparison of median filter, extremum median filter and the method in this paper is provided both in subjective images and objective MAE and MSE data. Obviously, the results indicated that the new method has better properties.

Signal processor, signal processing method and digital camera

Abstract: PROBLEM TO BE SOLVED: To provide a signal processor, a signal processing method and a digital camera capable of reducing noise while preventing the partial deterioration of image quality. SOLUTION: A microcomputer 40 detects a pixel located at an edge of an image displayed by an image signal outputted form a CCD 24 and controls a noise reduction processing part 29 so as not to perform noise reduction processing of pixels located at the edge and so as to perform noise reduction processing of the other pixels. COPYRIGHT: (C)2006,JPO&NCIPI

Image pickup apparatus capable of reducing noise in image signal and method for reducing noise in image signal

Abstract: Noise included in image data output from an analog processing unit, is eliminated by a noise reduction circuit. The noise reduction circuit restricts a difference between a value of each of peripheral pixels and a value of a center pixel to a value a value having a predetermined noise constant defined as an upper limit with respect to each of the peripheral pixels, e.g., a pixel space of 5×5. The values of the peripheral pixels whose upper limits are restricted are defined as new values of the peripheral pixels. The value of the center pixel is converted into a value obtained by averaging the new values of the peripheral pixels and the value of the center pixel. In this manner, low pass filter processing is carried out.

Restoration from Image Degraded by White Noise Based on Iterative Spectral Subtraction Method

Abstract: The application of the spectral subtraction method to image restoration is investigated. Comparison of the performance of the spectral subtraction method with that of the Wiener filter, in an ideal case, shows that the spectral subtraction method provides an improvement. In order to improve further the performance of the spectral subtraction method, an iterative algorithm is utilized with a frequency-division based noise estimation technique.

Elimination of quantization effects in measured temporal noise

Abstract: This paper presents a mathematical analysis of how temporal noise is transformed by quantization. A new method for measuring temporal noise with a low-resolution ADC and then accurately refer it back to the input of the ADC is shown. The method is, for instance, applicable to CMOS image sensors where photon shot noise is commonly used for determining conversion gain and quantum efficiency. Experimental tests have been carried out using a custom designed CMOS image sensor with an on-chip ADC featuring programmable gain and offset. The measurements verify the analysis and the method, e.g. noise levels of 0.11 LSB was measured with an accuracy 30 times higher than a traditional method would give.

Apparatus and method for measuring noise amount in video signal

Abstract: PROBLEM TO BE SOLVED: To solve a problem that it is difficult to accurately measure a noise amount because the noise and a motion reflect in a calculated value based on a difference between a preceding frame and a succeeding frame in the case of measuring the noise amount of a video signal within a video region. SOLUTION: An apparatus provided with a storage means for storing an input signal and a circuit for detecting the correlation measures the noise amount on the basis of the difference between images with high correlativity in the preceding frame and the succeeding frame. Since wrong measurement due to the motion can be prevented thereby, the noise measurement accuracy is improved. Further, since the noise amount can be measured with high accuracy, it is possible to properly adjust the noise reduction effect. COPYRIGHT: (C)2005,JPO&NCIPI

Noise estimation for remote sensing image data analysis

Abstract: Noise estimation does not receive much attention in remote sensing society. It may be because normally noise is not large enough to impair image analysis result. Noise estimation is also very challenging due to the randomness nature of the noise (for random noise) and the difficulty of separating the noise component from the signal in each specific location. We review and propose seven different types of methods to estimate noise variance and noise covariance matrix in a remotely sensed image. In the experiment, it is demonstrated that a good noise estimate can improve the performance of an algorithm via noise whitening if this algorithm assumes white noise.

Adaptive suppression of CCD signal-dependent noise in light space

Abstract: We propose a denoising scheme to restore images degraded by CCD noise. Typically, restoration algorithms assume a linear mapping between the incident light space and image space. However, in practice, a camera response function performs a non-linear mapping on the sensor output and, as a result, the sensor noise model becomes more complex in the image space. We correct for non-linearity by mapping the corrupted image into "light space", where the relationship between the incident light and light space values is linear. To reduce the sensor noise we accurately model the CCD sensor noise by using the photon transfer curve. We then develop a combination of adaptive filters based on the estimated noise model in light space. Our adaptive system demonstrates efficient noise removal performance in uniform regions, while preserving edges and fine details.

Noise modeling and estimation in image sequences from thermal infrared cameras

Abstract: In this paper we present an automated procedure devised to measure noise variance and correlation from a sequence, either temporal or spectral, of digitized images acquired by an incoherent imaging detector. The fundamental assumption is that the noise is signal-independent and stationary in each frame, but may be non-stationary across the sequence of frames. The idea is to detect areas within bivariate scatterplots of local statistics, corresponding to statistically homogeneous pixels. After that, the noise PDF, modeled as a parametric generalized Gaussian function, is estimated from homogeneous pixels. Results obtained applying the noise model to images taken by an IR camera operated in different environmental conditions are presented and discussed. They demonstrate that the noise is heavy-tailed (tails longer than those of a Gaussian PDF) and spatially autocorrelated. Temporal correlation has been investigated as well and found to depend on the frame rate and, by a small extent, on the wavelength of the thermal radiation.

Method, apparatus and program for noise suppression

Abstract: A band-limited image signal generating unit decomposes an input image signal into a plurality of band-limited image signals, each representing an image having different frequencies. An index value obtaining unit obtains an index value indicating a level of noise suppression based on data representing spatial frequencies as well as an evaluation value representing local contrast at a pixel of interest in band-limited images and data representing an X-ray dose. A noise suppression processing unit performs noise suppression processing on each pixel of each of band-limited images based on the index value. A processed image generating unit obtains a processed image, in which noise has been suppressed, by adding signals obtained by subtracting processed band-limited image signals, of which noise has been suppressed, from the band-limited image signals together, extracting a noise image signal, and subtracting the noise image signal from the input image signal.

The Analysis of Typical Method of Digital Image Noise Processing

Abstract: This paper introduces fundamental knowledge about digital image noise processing and illustrates several typical noise reduction methods such as neighbourhood averaging,median filtering and Wiener filtering, then it analyzes and compares the characteristics of them with the help of MATLAB software.

Noise in imaging systems: fixed pattern noise, electronic, and interference noise

Abstract: Imaging digital systems are widely used nowadays. CCD and CMOS sensors are embedded in a lot of metrologic devices for metrology in a lot of devices. One of these applications is the characterization of laser beams. For these kinds of applications, it is necessary to use cameras with high dynamic range. Some algorithms have been proposed in the past for this purpose. But, normally they enhance not only the dynamic range but the noise sensor too. In this paper we have applied an automatic algorithm to classify the different noise processes appearing in a CCD matrix. The method is based on a principal component expansion of the covariance matrix of some frames taken with the camera. It is possible to classify not only non-uniformity noise of the detector matrix, but also those contributions due to electronics and electronic interference and vibrations. Some of these noise processes represent only a very low amount of the total noise. A method to filter these noises is also presented.

Analysis on noises of image sensor of welding robot

Abstract: For the non-linearity of the reflectance of laser stripe on the work piece and performance of camera lens, the random noise added in the process of photoelectricity conversion, the storage, transmission and export of the electric charge of signal having the noises such as dark current, the video signal produced noises when transmitted, quantification noises hade by the A/D conversion of the video signal,definite error of the digital image signal will be resulted in the processes of acquiring the weld's information of the computer vision system of welding robot. In order to reduce the impact on the errors, the image noise of the image sensor is analyzed,and an experiential image noise model have been set up, the experimental results prove this model is effectual and applied.