Batch Normalization class. Subclasses the Keras BN class and hardcodes training=False so the BN layer doesn't update during training. Bat

Implements ROI Pooling on multiple levels of the feature pyramid. Params: - pool_shape: [height, width] of the output pooled regions. Usually

Configuration for training on MS COCO. Derives from the base Config class and overrides values specific to the COCO dataset.

Takes classified proposal boxes and their bounding box deltas and returns the final detection boxes. Returns: [batch, num_detections, (y1

Subsamples proposals and generates target box refinment, class_ids, and masks for each. Inputs: proposals: [batch, N, (y1, x1, y2, x2)] i

Receives anchor scores and selects a subset to pass as proposals to the second stage. Filtering is done based on anchor scores and non-max sup

* \brief Abstract interface for allocating HyperGraphElement */

* \brief base for allocating an optimization algorithm * * Allocating a solver for a given optimizer. The method construct() has to be * impl

* \brief Abstract interface for allocating a robust kernel */

forward declaration

* \brief base for the block solvers with some basic function interfaces */

Vector of words to represent images

Base configuration class. For custom configurations, create a sub-class that inherits from this one and override properties that need to be ch

The base class for dataset classes. To use it, create a new class that adds functions specific to the dataset you want to use. For example:

/

* \brief 7D edge between two Vertex7 */

/

* \brief propagation of an initial guess */

* \brief cost for traversing along active edges in the optimizer * * You may derive an own one, if necessary. The default is to return initialEs

* \brief cost for traversing only odometry edges. * * Initialize your graph along odometry edges. An odometry edge is assumed to connect vertice

Generic class to encapsulate functions to manage descriptors. * This class must be inherited. Derived classes can be used as the * parameter F when c

Functions to manipulate ORB descriptors

* \brief create vertices and edges based on TAGs in, for example, a file */

Vector of nodes with indexes of local features

* \brief statistics about the optimization */

Base class of scoring functions

Class that models a directed Hyper-Graph. An hyper graph is a graph where an edge can connect one or more nodes. Both Vertices and Edges of an h

* \brief Abstract action that operates on an entire graph */

* \brief library of actions, indexed by the action name; * * library of actions, indexed by the action name; * one can use ti to register a c

* \brief Abstract action that operates on a graph entity */

* \brief collection of actions * * collection of actions calls contains homogeneous actions operating on different types * all collected acti

THIS IS THE INITIALIZER FOR MONOCULAR SLAM. NOT USED IN THE STEREO OR RGBD CASE.