*Fua, P.[Pascal]*, and
*Hanson, A.J.*,

**Objective Functions for Feature Discrimination**,

*IJCAI89*(1596-1602). Other versions are:
BibRef
**8900**

**Objective Functions for Feature Discrimination: Theory**,

*DARPA89*(443-460).
BibRef

And:

**Objective Functions for Feature Discrimination:
Applications to Semiautomated and Automated Feature Extraction**,

*DARPA89*(676-694).
(SRI) Together, these papers use a
class of objective functions to order labels and maximize the probability of
correct division. It combines geometry (rectangles) and image data
(uniformity) to fit the shape to the data.
BibRef

*Fua, P.*, and
*Hanson, A.J.*,

**An Optimization Framework for Feature Extraction**,

*MVA(4)*, No. 2, Spring 1991, pp. 59-87.
BibRef
**9100**

*Fua, P.*, and
*Hanson, A.J.*,

**Extracting Generic Shapes Using Model-Driven Optimization**,

*DARPA88*(994-1004). Extension of last year's work:
BibRef
**8800**

**Using Generic Geometric Models for Intelligent Shape Extraction**,

*AAAI-87*(706-711).
BibRef

And:
*DARPA87*(227-233).
*Segmentation, Shapes*.
*Active Contours*.
*Building Recognition*.
*Road Detection*. Somewhat an extension of the earlier work. The different generic
region or boundary types are used in the extraction and refinement
of the segmentation. Applied to buildings, roads, and trees.
Pieces of the boundary are used to estimate locations for other
pieces.
BibRef

*Fua, P.[Pascal]*, and
*Hanson, A.J.*,

**Resegmentation Using Generic Shape:
Locating General Cultural Objects**,

*PRL(5)*, 1987, pp. 243-252.
BibRef
**8700**

Earlier:

**Locating Cultural Regions in Aerial Imagery using Geometric Cues**,

*DARPA85*(271-278).
*Segmentation, Edges*. (SRI) Improve the segmentations by using straight line sides of
regions as constraints. The regions are found by an Ohlander-type
segmentor and refined by finding straight lines as sides and
looking at the arrangement of the straight lines.
BibRef

*Fua, P.*, and
*Leclerc, Y.G.*,

**Model Driven Edge Detection**,

*MVA(3)*, No. 1, 1990, pp. 45-56.

PS File.
BibRef
**9000**

Earlier:
*DARPA88*(1016-1021).
BibRef

And: A2, A1:

**Finding Object Boundaries Using Guided Gradient Ascent**,

*DARPA87*(888-891).
*Segmentation, Edges*. Basically a simple large scale edge
detector where the shape of the edge is changed to fit the actual data.
BibRef

*Leclerc, Y.G.*,

**Constructing Simple Stable Descriptions for Image Partitioning**,

*IJCV(3)*, No. 1. May 1989, pp. 73-102.

Springer DOI

PS File.
BibRef
**8905**

Earlier:
*DARPA88*(365-382).
BibRef

*Leclerc, Y.G.*,

**Image and Boundary Segmentation Via Minimal-Length
Encoding on the Connection Machine**,

*DARPA89*(1056-1069).
Region segmentation and boundary encoding as segments.
BibRef
**8900**

Chapter on 2-D Region Segmentation Techniques, Snakes, Active Contours continues in

Combining Region and Edge Based Techniques .

Last update:Jun 13, 2019 at 09:53:00