# US 12,762,794 B2 · Energy and compute optimization of real time image converter using 2D to 3D rendering

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

- Number: US 12,762,794 B2
- Document ID: US12762794B2
- Status: Granted
- Kind: B2
- Application number: 18/759,475
- Filed: 2024-06-28
- Published: 2026-01-01
- Granted: 2026-10-13
- Earliest claimed: 2024-06-28
- Assignees: Dell Products L.P.
- Applicants: Dell Products L.P.
- Inventors: Qing Ye; Zhisong Liu; Rowland Shaw
- Web page: https://patentgrep.com/patent/US12762794B2

## CPC

- G06T 17/00: Three-dimensional [3D] modelling for computer graphics
- G06T 15/04: Texture mapping
- G06V 10/25: Determination of region of interest [ROI] or a volume of interest [VOI]
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

An image conversion device includes an image capture device and a renderer. The image capture device captures a plurality of two-dimensional (2D) images. The renderer receives the 2D images and renders a 3D model of an object captured in the 2D images. In rendering the 3D model, the renderer first renders a binary edge map of the object, and next models textures for the 3D model.

## Claims

### Claim 1

An image conversion device, comprising:
a hardware image capture device configured to capture a plurality of two-dimensional (2D) images; and
a hardware renderer configured to receive the 2D images and to render a three-dimensional (3D) model of an object captured in the 2D images, wherein in rendering the 3D model, the hardware renderer first renders a binary edge map of the object, and next models textures for the 3D model;
wherein in rendering the binary edge map, the hardware renderer minimizes a cross entropy (L) from the 2D images as:
L
=
∑
r
∈
R

C
⁡
(
r
)
·
log
[
α
⁡
(
C
~
(
r
)
)
]
+
(
1
-
C
⁡
(
r
)
)
·
log
[
1
-
α
[
C
~
(
r
)
]
]

2
,
where R is a set of rays in each batch of 2D images, α[{tilde over (C)}(r)] and C(r) are predicted and ground truth RGB (red, green, blue) colors for ray r, and α is a sigmoid function to map a binary value (i.e., to zero (0) or one (1) [0,1]).

### Claim 2 (depends on 1)

The image conversion device of claim 1, wherein the binary edge map includes contours and edges of the object.

### Claim 3 (depends on 1)

The image conversion device of claim 1, wherein, in modeling the textures for the 3D model, the hardware renderer further recalculates the binary cross entropy (L).

### Claim 4 (depends on 3)

The image conversion device of claim 3, wherein, in recalculating the binary cross entropy (L), the hardware renderer recalculates the binary cross entropy as:
L=ρr∈R∥C(r)−fθ[r,E(r)]∥2, where fθ is a mapping function with learnable parameters θ that takes both camera ray r and edge map E(r) to learn color information.

### Claim 5 (depends on 1)

The image conversion device of claim 1, wherein, in modeling the textures for the 3D model, the hardware renderer further recalculates the binary cross entropy (L).

### Claim 6 (depends on 5)

The image conversion device of claim 5, wherein, in recalculating the binary cross entropy (L), the hardware renderer recalculates the binary cross entropy as:
L=ρr∈R∥C(r)−fθ[r,E(r)]∥2, where fθ is a mapping function with learnable parameters θ that takes both camera ray r and edge map E(r) to learn color information.

### Claim 7 (depends on 1)

The image conversion device of claim 1, wherein, in modeling the textures for the 3D model, the hardware renderer further selects at least one Region of Interest for the object.

### Claim 8 (depends on 1)

The image conversion device of claim 1, further comprising:
a data storage device including a repository to store the 3D model.

### Claim 9

A method, comprising:
providing, in an image conversion device, an image capture device;
capturing, by the image capture device, a plurality of two-dimensional (2D) images;
providing, in the image conversion device, a renderer;
receiving, by the renderer, the 2D images; and
rendering, by the renderer, a three-dimensional (3D) model of an object captured in the 2D images;
wherein in rendering the 3D model, the renderer first renders a binary edge map of the object, and next models textures for the 3D model and in rendering the binary edge map, the renderer minimizes a binary cross entropy (L) from the 2D images as:
L
=
∑
r
∈
R

C
⁡
(
r
)
·
log
[
α
⁡
(
C
~
(
r
)
)
]
+
(
1
-
C
⁡
(
r
)
)
·
log
[
1
-
α
[
C
~
(
r
)
]

2
,
where R is a set of rays in each batch of images, α[{tilde over (C)}(r)] and C(r) are predicted and ground truth RGB (red, green, blue) colors for ray r, and α is a sigmoid function to map a binary value (i.e., to zero (0) or one (1) [0,1]).

### Claim 10 (depends on 9)

The method of claim 9, wherein the binary edge map includes contours and edges of the object.

### Claim 11 (depends on 9)

The method of claim 9, wherein, in modeling the textures for the 3D model, the renderer further recalculates the binary cross entropy (L).

### Claim 12 (depends on 11)

The method of claim 11, wherein, in recalculating the binary cross entropy (L), the renderer recalculates the binary cross entropy as:
L=ρr∈R∥C(r)−fθ[r,E(r)]∥2, where fθ is a mapping function with learnable parameters θ that takes both camera ray r and edge map E(r) to learn color information.

### Claim 13 (depends on 9)

The method of claim 9, wherein, in modeling the textures for the 3D model, the renderer further recalculates the binary cross entropy (L).

### Claim 14 (depends on 13)

The method of claim 13, wherein, in recalculating the binary cross entropy (L), the renderer recalculates the binary cross entropy as:
L=ρr∈R∥C(r)−fθ[r,E(r)]∥2, where fθ is a mapping function with learnable parameters θ that takes both camera ray r and edge map E(r) to learn color information.

### Claim 15 (depends on 9)

The method of claim 9, wherein, in modeling the textures for the 3D model, the renderer further selects at least one Region of Interest for the object.

### Claim 16

An image conversion device, comprising:
a hardware image capture device configured to capture a plurality of two-dimensional (2D) images;
a hardware renderer configured to receive the 2D images and to render a three-dimensional (3D) model of an object captured in the 2D images, wherein in rendering the 3D model, the hardware renderer first renders a binary edge map of the object, and next models textures for the 3D model; and
a hardware data storage device including a repository to store the 3D model;
wherein in rendering the binary edge map, the hardware renderer minimizes a cross entropy (L) from the 2D images, and in modeling the textures for the 3D model, the hardware renderer further recalculates the binary cross entropy (L) as:
L
=
∑
r
∈
R

C
⁡
(
r
)
·
log
[
α
⁡
(
C
~
(
r
)
)
]
+
(
1
-
C
⁡
(
r
)
)
·
log
[
1
-
α
[
C
~
(
r
)
]

2
,
where R is a set of rays in each batch of 2D images, α[{tilde over (C)}(r)] and C(r) are predicted and ground truth RGB (red, green, blue) colors for ray r, and α is a sigmoid function to map a binary value (i.e., to zero (0) or one (1) [0,1]).