Struct vlfeat_sys::_VlAIB
[−]
[src]
#[repr(C)]pub struct _VlAIB { pub nodes: *mut vl_uint, pub nentries: vl_uint, pub beta: *mut f64, pub bidx: *mut vl_uint, pub which: *mut vl_uint, pub nwhich: vl_uint, pub Pcx: *mut f64, pub Px: *mut f64, pub Pc: *mut f64, pub nvalues: vl_uint, pub nlabels: vl_uint, pub parents: *mut vl_uint, pub costs: *mut f64, pub verbosity: vl_uint, }
** @internal ** @brief AIB algorithm data ** ** The implementation is quite straightforward, but the way feature ** values are handled in order to support efficient joins, ** deletions and re-arrangement needs to be explained. This is ** achieved by adding a layer of indirection: ** - Call each feature value (either original or obtained by a join ** operation) a node. Nodes are identified by numbers. ** - Call each element of the various arrays (such as VlAIB::Px) ** an entry. ** - Entries are dynamically associated to nodes as specified by ** VlAIB::nodes. For example, @c Px[i] refers to the node @c ** nodes[i].
Fields
nodes: *mut vl_uint
< Entires to nodes
nentries: vl_uint
< Total number of entries (= # active nodes)
beta: *mut f64
< Minimum distance to an entry
bidx: *mut vl_uint
< Closest entry
which: *mut vl_uint
< List of entries to update
nwhich: vl_uint
< Number of entries to update
Pcx: *mut f64
< Joint probability table
Px: *mut f64
< Marginal.
Pc: *mut f64
< Marginal.
nvalues: vl_uint
< Number of feature values
nlabels: vl_uint
< Number of labels
parents: *mut vl_uint
< Array of parents
costs: *mut f64
< Cost of each merge
verbosity: vl_uint
Trait Implementations
impl Debug for _VlAIB
[src]
impl Copy for _VlAIB
[src]
impl Clone for _VlAIB
[src]
fn clone(&self) -> Self
Returns a copy of the value. Read more
fn clone_from(&mut self, source: &Self)
1.0.0
Performs copy-assignment from source
. Read more