# US 2002/0054236 A1 · Method and apparatus for motion adaptive deinterlacing

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2002/0054236 A1
- Document ID: US20020054236A1
- Status: Published application
- Kind: A1
- Application number: 09/948,721
- Filed: 2001-09-10
- Published: 2002-05-09
- Earliest claimed: 2000-09-08
- Assignees: Pixelworks, Inc.
- Inventors: Finn G. Wredenhagen; Gary Cheng; Soo Jang; Lance Greggain
- Web page: https://patentgrep.com/patent/US20020054236A1

## CPC

- H04N 5/142: {Edging; Contouring}
- H04N 5/144: {Movement detection (for video coding H04N19/503; analysis of motion in general G06T7/20)}
- H04N 7/012: {Conversion between an interlaced and a progressive signal (for graphics images G09G2310/0229)}
- H04N 7/0137: {dependent on presence/absence of motion, e.g. of motion zones (H04N7/014 takes precedence; movement detection in television signals H04N5/144)}
- H04N 19/117: Filters, e.g. for pre-processing or post-processing (sub-band filter banks H04N19/635)
- H04N 19/139: Analysis of motion vectors, e.g. their magnitude, direction, variance or reliability
- H04N 19/51: Motion estimation or motion compensation
- H04N 7/0115: {with details on the detection of a particular field or frame pattern in the incoming video signal, e.g. 3:2 pull-down pattern}
- H04N 7/0142: {the interpolation being edge adaptive (edge detection in television signals H04N5/142; edge-driven scaling G06T3/403)}
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION

## Abstract

An adaptive filter calculates a target pixel from an interlaced video signal. The video signal comprises a plurality of frames, each of which comprise an even and an odd field. The filter comprises a quantized motion calculator and a filter selector. The quantized motion calculator estimates an amount of motion about the target pixel. The filter selector selects a filter in accordance with the estimated amount of motion. The filter applies a first weighting factor to a plurality of current field pixels and a second weighting factor to a plurality of previous field pixels for creating the target pixel.

## Claims

### Claim 1

An adaptive filter for calculating a target pixel from an interlaced video signal comprising a plurality of frames, each of said frames comprising an even and an odd field, said filter comprising:
a) a quantized motion calculator for estimating an amount of motion about said target pixel; and
b) a filter selector for selecting a filter in accordance with said estimated amount of motion, wherein said filter applies a first weighting factor to a plurality of current field pixels and a second weighting factor to a plurality of previous field pixels for creating said target pixel.

### Claim 2 (depends on 1)

An adaptive filter as defined in claim 1, wherein said quantized motion calculator estimates said amount of motion about said target pixel by calculating a difference between a first pixel and a second pixel, said first and second pixels being spatially equivalent and temporally different by one frame, said difference being quantized in accordance with a plurality of predefined thresholds.

### Claim 3 (depends on 2)

An adaptive filter as defined in claim 2, wherein if said amount of motion of said first pixel is deemed static, a corresponding value is provided for indicating the number of consecutive frames said first pixel location has been static.

### Claim 4 (depends on 3)

An adaptive filter as defined in claim 3, wherein said corresponding value appended to said target pixel's information.

### Claim 5 (depends on 1)

An adaptive filter as defined in claim 1, said filter further comprising a feathering detector for detecting a feathering artifact about said target pixel, wherein said feathering detection enhances reliability of said motion detection.

### Claim 6 (depends on 5)

An adaptive filter as defined in claim 5, wherein said feathering detector detects feathering if each of a plurality of pixel differences along a predefined contour exceeds corresponding predefined thresholds.

### Claim 7 (depends on 6)

An adaptive filter as defined in claim 6, wherein said predefined contour is selected from a plurality of contours, including non-linear contours.

### Claim 8 (depends on 6)

An adaptive filter as defined in claim 6, wherein said predefined thresholds are independent of each other.

### Claim 9 (depends on 6)

An adaptive filter as defined in claim 6, wherein said feathering detector detects feathering if each of said plurality of pixel differences exceeds said corresponding predefined thresholds for a predefined number of consecutive fields.

### Claim 10 (depends on 9)

An adaptive filter as defined in claim 9, wherein feathering is detected concurrently over a plurality of different values for said predefined number of consecutive fields.

### Claim 11 (depends on 6)

An adaptive filter as defined in claim 6, wherein a corresponding value indicating whether or not feathering is present is appended to said target pixel's information.

### Claim 12 (depends on 1)

An adaptive filter as defined in claim 1, said filter further comprising an edge detector for detecting static edges about said target pixel, wherein said edge detection enhances the reliability of said motion detection.

### Claim 13 (depends on 12)

An adaptive filter as defined in claim 12, wherein said edge detector detects spatio-temporal edges in said even and odd fields by calculating a difference between pixels in said even and odd fields.

### Claim 14 (depends on 13)

An adaptive filter as defined in claim 13, wherein if an edge is detected, a corresponding value is provided for indicating the number of consecutive fields said edge has been present.

### Claim 15 (depends on 14)

An adaptive filter as defined in claim 14, wherein said corresponding value is appended to said target pixel's information.

### Claim 16 (depends on 12)

An adaptive filter as defined in claim 12, wherein said edge detector detects an edge structure by looking for an edge followed by a flat portion followed by another edge.

### Claim 17 (depends on 1)

An adaptive filter as defined in claim 1, wherein said filter is applied to a contour of current and previous field pixels for interpolating a target pixel.

### Claim 18 (depends on 17)

An adaptive filter as defined in claim 17, wherein said contour is selected from a plurality of contours, including non-linear contours.

### Claim 19

A feathering detector for detecting a feathering artifact in adjacent odd and even fields about a target pixel, said feathering detector comprising:
(a) a contour selector for selecting from a plurality of contours, said contours including non-linear contours;
(b) a difference calculator for calculating a plurality of differences between pixels along said selected contour; and
(c) a plurality of predefined thresholds for comparing with said calculated differences for determining if feathering exists about said pixel, whereby including non-linear contours improves feathering detection in arbitrary shapes.

### Claim 20 (depends on 19)

A feathering detector as defined in claim 19, wherein said predefined thresholds are programmable.

### Claim 21 (depends on 19)

A feathering detector as defined in claim 19, wherein some of said predefined thresholds are different from each other for further enhancing the detection of arbitrary shapes.

### Claim 22 (depends on 19)

A feathering detector as defined in claim 19, wherein feathering exits only if feathering is detected for a predefined number of consecutive pixels, whereby said predefined number of pixels is taken as a group.

### Claim 23 (depends on 22)

A feathering detector as defined in claim 22, wherein feathering is detected concurrently over a plurality of different values for said predefined number of consecutive pixels.

### Claim 24

An adaptive filter for calculating a target pixel from an interlaced video signal comprising a plurality of frames, each of said frames comprising an even and an odd field, said filter comprising:
a) a quantized motion calculator for estimating an amount of motion between consecutive frames about said target pixel;
b) a motion structure detector for detecting feathering between consecutive fields, wherein said feathering is indicative of motion;
c) an edge structure detector for detecting vertical edges about said target pixel, wherein said edge is indicative of an absence of motion;
d) a static structure detection block for detecting patterns indicative of motion and patterns indicative of no motion; and
e) a filter selector for selecting a filter in accordance with said estimated amount of motion, wherein said filter applies a first weighting factor to a plurality of current field pixels and a second weighting factor to a plurality of previous field pixels for creating said target pixel.

### Claim 25 (depends on 24)

An adaptive filter as defined in claim 24, wherein at least a portion of information calculated by said quantized motion calculator, motion structure detector, edge structure detector, and static structure detector is appended to said target pixel information.

### Claim 26

An method for selecting a filter for calculating a target pixel from an interlaced video signal comprising a plurality of frames, each of said frames comprising an even and an odd field, said method comprising the steps of:
a) estimating an amount of motion between consecutive frames about said target pixel;
b) detecting feathering between consecutive fields, wherein said feathering is indicative of motion;
c) detecting vertical edges about said target pixel, wherein said edge is indicative of an absence of motion;
d) detecting patterns indicative of motion and patterns indicative of no motion;
e) adjusting said estimated amount motion in accordance with said feathering, edges and pattern detection; and
f) selecting a filter in accordance with said estimated amount of motion and applying a first weighting factor to a plurality of current field pixels and a second weighting factor to a plurality of previous field pixels for creating said target pixel.
