cecile-supercool-tracker 0.0.1

Machine learning framework for building object trackers and similarity search engines
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167

use crate::utils::kalman::kalman_2d_point::{Point2DKalmanFilter, DIM_2D_POINT_X2};
use crate::utils::kalman::KalmanState;
use nalgebra::Point2;

#[derive(Debug)]
pub struct Vec2DKalmanFilter {
    f: Point2DKalmanFilter,
}

/// Default initializer
impl Default for Vec2DKalmanFilter {
    fn default() -> Self {
        Self {
            f: Point2DKalmanFilter::new(1.0 / 20.0, 1.0 / 160.0),
        }
    }
}

impl Vec2DKalmanFilter {
    pub fn new(position_weight: f32, velocity_weight: f32) -> Self {
        Self {
            f: Point2DKalmanFilter::new(position_weight, velocity_weight),
        }
    }

    pub fn initiate(&self, points: &[Point2<f32>]) -> Vec<KalmanState<DIM_2D_POINT_X2>> {
        points.iter().map(|p| self.f.initiate(p)).collect()
    }

    pub fn predict(
        &self,
        state: &[KalmanState<DIM_2D_POINT_X2>],
    ) -> Vec<KalmanState<DIM_2D_POINT_X2>> {
        state.iter().map(|s| self.f.predict(s)).collect()
    }

    pub fn update(
        &self,
        state: &[KalmanState<DIM_2D_POINT_X2>],
        points: &[Point2<f32>],
    ) -> Vec<KalmanState<DIM_2D_POINT_X2>> {
        assert_eq!(
            state.len(),
            points.len(),
            "Lengths of state and points must match"
        );
        state
            .iter()
            .zip(points.iter())
            .map(|(s, p)| self.f.update(s, p))
            .collect()
    }

    pub fn distance(
        &self,
        state: &[KalmanState<DIM_2D_POINT_X2>],
        points: &[Point2<f32>],
    ) -> Vec<f32> {
        assert_eq!(
            state.len(),
            points.len(),
            "Lengths of state and points must match"
        );
        state
            .iter()
            .zip(points.iter())
            .map(|(s, p)| self.f.distance(s, p))
            .collect()
    }

    pub fn calculate_cost(distances: &[f32], inverted: bool) -> Vec<f32> {
        distances
            .iter()
            .map(|d| Point2DKalmanFilter::calculate_cost(*d, inverted))
            .collect()
    }
}

#[cfg(feature = "python")]
pub mod python {
    use crate::utils::kalman::kalman_2d_point::python::PyPoint2DKalmanFilterState;
    use crate::utils::kalman::kalman_2d_point_vec::Vec2DKalmanFilter;
    use nalgebra::Point2;
    use pyo3::prelude::*;

    #[pyclass]
    #[pyo3(name = "Vec2DKalmanFilter")]
    pub struct PyVec2DKalmanFilter {
        filter: Vec2DKalmanFilter,
    }

    #[pymethods]
    impl PyVec2DKalmanFilter {
        #[new]
        #[pyo3(signature = (position_weight = 0.05, velocity_weight = 0.00625))]
        pub fn new(position_weight: f32, velocity_weight: f32) -> Self {
            Self {
                filter: Vec2DKalmanFilter::new(position_weight, velocity_weight),
            }
        }

        #[pyo3(signature = (points))]
        pub fn initiate(&self, points: Vec<(f32, f32)>) -> Vec<PyPoint2DKalmanFilterState> {
            let args = points
                .iter()
                .map(|(x, y)| Point2::from([*x, *y]))
                .collect::<Vec<_>>();

            self.filter
                .initiate(&args)
                .into_iter()
                .map(PyPoint2DKalmanFilterState::new)
                .collect()
        }

        #[pyo3(signature = (state))]
        pub fn predict(
            &self,
            state: Vec<PyPoint2DKalmanFilterState>,
        ) -> Vec<PyPoint2DKalmanFilterState> {
            let args = state.into_iter().map(|s| *s.inner()).collect::<Vec<_>>();
            self.filter
                .predict(&args)
                .into_iter()
                .map(PyPoint2DKalmanFilterState::new)
                .collect()
        }

        #[pyo3(signature = (state, points))]
        pub fn update(
            &self,
            state: Vec<PyPoint2DKalmanFilterState>,
            points: Vec<(f32, f32)>,
        ) -> Vec<PyPoint2DKalmanFilterState> {
            let point_args = points
                .iter()
                .map(|(x, y)| Point2::from([*x, *y]))
                .collect::<Vec<_>>();
            let state_args = state.iter().map(|s| *s.inner()).collect::<Vec<_>>();
            self.filter
                .update(&state_args, &point_args)
                .into_iter()
                .map(PyPoint2DKalmanFilterState::new)
                .collect()
        }

        #[pyo3(signature = (state, points))]
        pub fn distance(
            &self,
            state: Vec<PyPoint2DKalmanFilterState>,
            points: Vec<(f32, f32)>,
        ) -> Vec<f32> {
            let point_args = points
                .iter()
                .map(|(x, y)| Point2::from([*x, *y]))
                .collect::<Vec<_>>();
            let state_args = state.iter().map(|s| *s.inner()).collect::<Vec<_>>();
            self.filter.distance(&state_args, &point_args)
        }

        #[staticmethod]
        #[pyo3(signature = (distances, inverted))]
        pub fn calculate_cost(distances: Vec<f32>, inverted: bool) -> Vec<f32> {
            Vec2DKalmanFilter::calculate_cost(&distances, inverted)
        }
    }
}