In this paper, a technique is presented to alleviate ghosting artifacts in the decoded video sequences for low-bit-rate
video coding. Ghosting artifacts can be defined as the appearance of ghost like outlines of an object in a decoded video
frame. Ghosting artifacts result from the use of a prediction loop in the video codec, which is typically used to increase
the coding efficiency of the video sequence. They appear in the presence of significant frame-to-frame motion in the
video sequence, and are typically visible for several frames until they eventually die out or an intra-frame refresh occurs.
Ghosting artifacts are particularly annoying at low bit rates since the extreme loss of information tends to accentuate
their appearance. To mitigate this effect, a procedure with selective in-loop filtering based on motion vector information
is proposed. In the proposed scheme, the in-loop filter is applied only to the regions where there is motion. This is done
so as not to affect the regions that are devoid of motion, since ghosting artifacts only occur in high-motion regions. It is
shown that the proposed selective filtering method dramatically reduces ghosting artifacts in a wide variety of video
sequences with pronounced frame-to-frame motion, without degrading the motionless regions.
In this paper, we present a region-of-interest-based video coding system for use in real-time applications. Region-of-interest (ROI) coding methodology specifies that targets or ROIs be coded at higher fidelity using a greater number of available bits, while the remainder of the scene or background is coded using a fewer number of bits. This allows the target regions within the scene to be well preserved, while dramatically reducing the number of bits required to code the video sequence, thus reducing the transmission bandwidth and storage requirements. In the proposed system, the ROI contours can be selected arbitrarily by the user via a graphical user interface (GUI), or they can be specified via a text file interface by an automated process such as a detection/tracking algorithm. Additionally, these contours can be specified at either the transmitter or receiver. Contour information is efficiently exchanged between the transmitter and receiver and can be adjusted on the fly and in real time. Coding results are presented for both electro-optical (EO) and infrared (IR) video sequences to demonstrate the performance of the proposed system.
In this paper, we present a memory-efficient, contour-based, region-of-interest (ROI) algorithm designed for ultra-low-bit-
rate compression of very large images. The proposed technique is integrated into a user-interactive wavelet-based
image coding system in which multiple ROIs of any shape and size can be selected and coded efficiently. The coding
technique compresses region-of-interest and background (non-ROI) information independently by allocating more bits to
the selected targets and fewer bits to the background data. This allows the user to transmit large images at very low
bandwidths with lossy/lossless ROI coding, while preserving the background content to a certain level for contextual
purposes. Extremely large images (e.g., 65000 X 65000 pixels) with multiple large ROIs can be coded with minimal
memory usage by using intelligent ROI tiling techniques. The foreground information at the encoder/decoder is
independently extracted for each tile without adding extra ROI side information to the bit stream. The arbitrary ROI
contour is down-sampled and differential chain coded (DCC) for efficient transmission. ROI wavelet masks for each tile
are generated and processed independently to handle any size image and any shape/size of overlapping ROIs. The
resulting system dramatically reduces the data storage and transmission bandwidth requirements for large digital images
with multiple ROIs.