Eye-tracking eyewear
By using infrared radiation sources and detectors in smart eye-wearing devices, combined with orthogonal modulation and time-division multiplexing techniques, and incorporating artificial neural networks, the problems of large size and insufficient accuracy of existing eye-tracking systems have been solved, achieving high-precision and low-invasive eye-tracking.
Patent Information
- Application Number
- CN202580027841.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2024-06-07
- Filing Date
- 2025-05-23
- Publication Date
- 2026-11-17
AI Technical Summary
Existing eye-tracking systems are bulky in smart eyewear, unlike regular eyewear, and lack accuracy.
By employing at least two infrared radiation sources and infrared detectors, combined with a processing and control unit, and through noise filtering, signal modulation and demodulation, and utilizing orthogonal modulation signals and time-division multiplexing technology, efficient detection of infrared light and maximization of signal-to-noise ratio are achieved. The gaze direction is estimated by combining artificial neural networks.
It achieves high-precision eye tracking in smart eyewear, reduces infrared light exposure to the eyes, lowers invasiveness, and improves the accuracy and safety of the system.
Smart Images

Figure CN123125530A_ABST
Abstract
Description
[0002] This invention relates to eye-wearing devices, and more particularly to smart eye-wearing devices including an eye-tracking system.
[0003] Smart eyewear encompasses any type of device designed to be worn on at least one of the wearer's eyes and having a frame configured to be worn on the head, including any headband device.
[0004] As is known, smart eye wearables include electronic devices adapted to function as electronic calculators and capable of supporting activities performed by users wearing these devices.
[0005] More specifically, electronic devices can, for example, play audio and video, launch and run software applications, and recognize shapes and objects placed in their surrounding environment.
[0006] More specifically, the electronic devices typically include processing and control units, as well as sensors, integrated into the eyeglass frames or temples.
[0007] For example, eye-tracking systems are often provided to smart eye-wearing devices to enable / disable some of the device's functions, such as for augmented and mixed reality, and / or to monitor eye movement for medical, psychological, commercial, and marketing purposes, such as tracking the direction of gaze toward a store window or supermarket shelf.
[0008] Several different types of eye-tracking systems are currently known.
[0009] The most widely used first-type eye-tracking system consists of one or more sources for illuminating the eye being tracked and one or more cameras for capturing the movement of the eye. However, such systems are bulky and therefore cannot be embedded in smart eyewear that must look the same as ordinary eyewear.
[0010] Given this drawback, a second type of eye-tracking system becomes more attractive, comprising one or more infrared and / or near-infrared wavelength sources illuminating the eye to be tracked and a detector for detecting the optical signal reflected by the illuminated eye; in these systems, the gaze direction is obtained based on the detected optical signal. Because these systems operate at infrared and / or near-infrared wavelengths outside the sensitivity range of the human eye, they do not interfere with visual abilities.
[0011] In this specification, infrared wavelength is intended to indicate a wavelength between 700 nm and 1 mm.
[0012] Near-infrared wavelengths are intended to indicate wavelengths between 700 nm and 10 µm.
[0013] US 5,345,281, US 2018 / 0113508 A1, and US 2019 / 0384388 A1 describe examples of the second type of eye-tracking system.
[0014] The purpose of this invention is to design an eye-wearable device with a more accurate eye-tracking system than existing eye-tracking systems.
[0015] This and other objectives according to the invention are achieved by manufacturing an eye-wearing device with an eye-tracking system as described in claim 1.
[0016] Other features of eye-wearing devices with eye-tracking systems are the subject of the dependent claims.
[0017] The characteristics and advantages of the eye-wearing device with an eye-tracking system according to the invention will become more apparent from the following exemplary but non-limiting description with reference to the accompanying schematic diagrams, in which:
[0018] - Figure 1 This is a schematic perspective view of an eye-wearing device according to the present invention;
[0019] - Figure 2 This is a block diagram illustrating the electronic device included in an eye-wearing device according to the present invention;
[0020] - Figure 3 This is a block diagram representing the modulation of an infrared radiation source by N square waves of different frequencies, the demodulation of signals detected by M infrared detectors, and the acquisition of data for each eye position. This block is shown as an M×N matrix.
[0021] Referring to the accompanying drawings, an eye-wearing device is shown, which is generally referred to as 100.
[0022] The eyewear 100 includes a frame 110, which may include a front frame 111, two temples 112, 113 connected to the front frame 111, and at least one lens 114, 115 connected to the front frame 111. For example, the eyewear is eyeglasses, and in this case, the eyewear may include two lenses; the eyewear 100 may also be a ski mask or a face mask-style sunglasses, and in this case, the eyewear may include a single lens.
[0023] The eyewear 100 advantageously includes an eye-tracking system 200 associated with a frame 110. The eye-tracking system 200 includes at least two infrared radiation sources 210 coupled to the frame 110 so as to face and emit infrared radiation toward the corresponding eye of the user to be tracked when the user wears the eyewear 100.
[0024] In the illustrated embodiment, at least two infrared radiation sources 210 are present. Specifically, the eye-tracking system 200 may include at least two infrared radiation sources 210 for each eye to be tracked. In any case, at least two infrared radiation sources 210 must be pointed at the eye to be tracked.
[0025] For example, at least two infrared radiation sources 210 are two LEDs.
[0026] The eye-tracking system 200 includes at least one infrared detector 220 attached to the frame 110.
[0027] Specifically, at least one infrared detector 220 may be attached to the front frame 111 or to at least one lens 114, 115.
[0028] At least one infrared detector 220 is configured to detect an optical signal reflected by the eye to be tracked, the optical signal being an infrared wavelength, preferably a near-infrared wavelength, more preferably a wavelength of about 850 nm; at least one infrared detector 220 is also configured to generate a corresponding electrical detection signal.
[0029] For example, at least one infrared detector 220 can be manufactured as a solid-state semiconductor detector, such as a photodiode. For example, the infrared detector 220 can be made of silicon or other semiconductors, or other materials (such as solution-processable materials).
[0030] The electrical detection signal generated by at least one infrared detector 220 can be an analog signal.
[0031] In particular, the electrical detection signal generated by at least one infrared detector 220 can be an analog current detection signal.
[0032] At least one infrared detector 220 may be directly or indirectly connected to an analog signal converter circuit 222, which converts an analog current detection signal into an analog voltage detection signal. The output of this analog signal converter circuit 222 may be connected to an analog-to-digital converter (ADC) 221, which is configured to convert the analog electrical detection signal into a digital electrical detection signal.
[0033] In some embodiments, there is no analog signal converter circuit 222, and at least one infrared detector 220 may be directly or indirectly connected to the ADC converter 221.
[0034] Furthermore, the analog signal converter circuit 222 (if present) and the ADC converter can be embedded in the detector 220, or they can be external relative to the detector 220. If the analog signal converter circuit 222 and the ADC converter 221, which may be present, are embedded in the detector 220, then the electrical detection signal generated by the infrared detector 220 is a digital signal.
[0035] The eye-tracking system 200 also includes a processing and control unit 230 associated with at least two infrared radiation sources 210 and at least one infrared detector 220. Specifically, the processing and control unit 230 is configured to control each of the at least two infrared radiation sources 210 to emit infrared radiation modulated with a corresponding modulation signal, receive an electrical detection signal, and estimate the gaze direction of the eye to be tracked based on the received electrical detection signal. More specifically, the modulation signals may have different frequencies and different shapes.
[0036] For example, the processing and control unit 230 is a microcontroller.
[0037] Preferably, the eye-tracking system 200 further includes a demodulation module 280 for each infrared radiation source 210. The demodulation module 280 is specifically associated with the at least one infrared detector 220.
[0038] Each demodulation module 280 is configured to filter the received electrical detection signal with a noise filter suitable for the corresponding infrared radiation source 210, wherein the noise filter is configured to reduce the noise contribution in the received electrical detection signal.
[0039] The filter can be analog or digital. In the case of an analog filter, the demodulation module 280 is connected to the output of at least one infrared detector 220.
[0040] Specifically, the output of demodulation module 280 can be directly connected to ADC converter 221, the output of which is directly connected to processing and control unit 230. Alternatively, the output of demodulation module 280 can be directly connected to analog signal converter circuit 222, the output of which is directly connected to ADC converter 221, and the output of which is directly connected to processing and control unit 230.
[0041] In contrast, when the filter is digital, the demodulation module 280 is integrated into the processing and control unit 230. In this case, the output of at least one infrared detector 220 can be directly connected to the analog signal converter circuit 222, whose output is directly connected to the ADC converter 221, whose output is directly connected to the processing and control unit 230. Alternatively, the output of at least one infrared detector 220 can be directly connected to the ADC converter 221, whose output is directly connected to the processing and control unit 230.
[0042] Preferably, the noise filter is configured to minimize the noise contribution in the received electrical detection signal; in this case, the noise filter is called a matched noise filter, and it has a time transfer function equal to the mirror time function of the modulation signal relative to the infrared radiation source 210.
[0043] A mirror-time function is defined as a function that has a shape that is a mirror image of the original function, where the mirror image occurs around a certain time.
[0044] Consider the following eye-tracking system 200, which has N infrared radiation sources 210 (where N ≥ 2) and M infrared detectors 220 (where M ≥ 1).
[0045] When N infrared radiation sources 210 emit modulated infrared radiation toward the eye to be tracked, this modulated infrared radiation is reflected, refracted, and scattered by different parts of the eye and the surrounding skin. Due to the rotation of the eyeball's azimuth and elevation angles and the movement of the eyelids (blinking), the light emitted by each of the N infrared radiation sources 210 is absorbed, reflected, and scattered in different directions and with different intensities. Each of M infrared detectors 220 receives the contribution of the modulated infrared radiation reflected and scattered by the eye and surrounding skin. Using a time function f... n The modulation signal (t) is used to modulate the intensity of infrared radiation emitted by each of the N infrared radiation sources 210, where the subscript n (ranging from 1 to N) indicates the nth source.
[0046] Under the assumption that the distance of light travel between N infrared radiation sources 210 and M infrared detectors 220 can be considered as a Gaussian channel and that ambient light noise can be considered as an additive white Gaussian process, the signal g received by the m-th infrared detector can be written as follows: m (t) equals:
[0047] g m (t) = a n,m f n (t) + w(t)
[0048] Among them, a n,m This represents the attenuation of the signal emitted by the nth infrared source and received by the mth infrared detector; this attenuation is caused by absorption by the eye and surrounding area, as well as dispersion caused by the absorption cone of the mth infrared detector. It is assumed that the attenuation is constant during the time period the signal is detected by the mth infrared detector. Therefore, the duration of each individual detection must be negligible compared to eye or eyelid movement.
[0049] h m (t) is defined as a filter applied to the signal detected by the m-th detector, and the value at the output of the filter can be defined as:
[0050]
[0051] Among them, symbols This represents the convolution operation, and This represents the white noise at the output of the noise filter.
[0052] Therefore, the signal-to-noise ratio at the output of the receiving filter at time τ can be defined as:
[0053]
[0054] Since the Fourier transform is a unitary operator, by applying Parseval's theorem, we can write:
[0055]
[0056] By applying Schwartz's inequality, we can write:
[0057]
[0058] The equation that maximizes SNR holds true is:
[0059]
[0060] Therefore, by applying the inverse Fourier transform, we can obtain... .
[0061] Then, the matched noise filter is equal to the mirror time function of the modulated signal, and sampling should be performed after the duration of the modulated infrared radiation has ended.
[0062] Therefore, time τ is the duration of the matched noise filter; furthermore, it represents the optimal moment for sampling so that the matched noise filter can be used in combination to maximize the SNR.
[0063] To minimize inter-symbol interference (ISI) and to distinguish the contribution of each infrared source n to each infrared detector m, when transmitting simultaneously from each infrared source, it is necessary for the infrared sources to use modulated signals that are orthogonal to each other, i.e.:
[0064]
[0065] This method enables each infrared detector to completely separate the contribution of each infrared source, thereby achieving maximum separability between signals.
[0066] Preferably, the modulated signals are substantially orthogonal to each other.
[0067] For example, the modulating signal is a sinusoidal signal with different frequencies. In order to satisfy the orthogonality condition of all N sources at each sampling time (a multiple of the period τ) reported above, the frequencies used for the different sinusoids must be multiples of the fundamental frequency 1 / τ.
[0068] However, the modulating signal can also be non-sinusoidal. For example, the modulating signal can be a square wave, which is easier to implement in embedded systems.
[0069] In the sinusoidal case, for each of the N sine waves, the value obtained by filtering each infrared detector corresponds to the Fourier series value for that specific frequency, that is, the period of the relative sinusoidal signal or a multiple of the period.
[0070] The previously reported equations refer to the case of an analog filter. In this case, the filtered signal is then sampled at a sampling period equal to a multiple of the duration of the matched noise filter (i.e., time τ). In the digital case, all the above equations must be transformed into the discrete time domain, and the entire demodulation process is performed by the demodulation module 280 incorporated in the processing and control unit 230.
[0071] In this case (digital filter), it is necessary to sample the signal from each infrared detector at a frequency at least twice the maximum frequency of the modulated sine wave in order to follow the sampling theorem.
[0072] This allows the amplitude of each infrared source received by each infrared detector to be determined directly using the Discrete Fourier Transform (DFT). The sampling frequency should also be a multiple of the maximum frequency sine wave frequency to avoid the effects of frequency leakage between different frequencies and to maximize the SNR.
[0073] Specifically, the processing and control unit 230 may include an internal memory (not shown) in which a computer program is loaded; such computer program includes instructions that, when executed by the electronic processing unit, actuate an estimation algorithm to estimate the gaze direction of the eye to be tracked based on the received electrical detection signal.
[0074] Specifically, at the end of the detection process of the M infrared detectors, a set of amplitude values of signals emitted by the N infrared sources and received by each of the M infrared detectors is obtained. These values are collected by the processing and control unit 230 into a single detection vector of length M × N elements, which can be considered as a vector in M × N dimensional space.
[0075] The processing and control unit 230 is then configured to associate the azimuth and elevation angles of the eyeball relative to the eyeglass frame with each detection vector. This association is performed by applying a nonlinear transformation function between an M × N dimensional space and a two-dimensional space represented by these two angles. This nonlinear transformation can be performed using various estimation algorithms that implement pattern recognition techniques, such as the following techniques:
[0076] - Lookup table interpolation, whose underlying data can be obtained from a series of real samples;
[0077] - Neural networks based on radial basis functions;
[0078] - Kohonen networks were developed specifically for nonlinear mappings from high-dimensional spaces to two-dimensional spaces.
[0079] Typically, other techniques based on artificial neural networks (ANNs) and nonlinear mappings can be used.
[0080] This processing allows the amount of IR radiation received from each infrared detector and emitted by each infrared source to be correlated with the eye's azimuth, elevation, and rotation angles. It also allows for the identification of eyelid closure (blinking).
[0081] In the case of an estimation algorithm based on an artificial neural network, the algorithm is trained using a dataset of appropriately sized detection signal maps associated with the corresponding gaze direction. The dataset can consist of simulated data generated with known gaze directions and experimental data. The former can be automatically acquired, for example, using an artificial eye that has a response similar to the human eye in the infrared spectrum, while the latter (experimental data) can be obtained by asking subjects to view a specific target in a specific direction relative to the frame of their eyewear. The artificial eye can be mounted on a motor-actuated platform, the position of which can be scanned using a dense grid of points identified by paired angles (specifically elevation and azimuth) defining the gaze direction. Training allows the weights of the ANN stored in internal memory to be determined for real-time eye-tracking estimation.
[0082] Therefore, the training process of the algorithm implemented by the computer program stored in the internal memory of the processing and control unit 230 includes the following steps:
[0083] - Perform the experimental dataset acquisition step, which includes the following steps:
[0084] - Provides an eye-wearing device 100 equipped with an eye-tracking system 200;
[0085] - Provide an artificial eye mounted on a support so that it can rotate about two rotation axes according to multiple attitudes defined by corresponding pitch or tilt angles and corresponding azimuth or yaw angles;
[0086] - Position the eyepiece 100 in front of the artificial eye;
[0087] - Scan a dense grid of points in angular space (i.e., composed of pairs of elevation and azimuth angles within the angular range to be estimated by the eye tracker system 200) and acquire electrical detection signals from at least one infrared detector 220 for each point;
[0088] - The algorithm is trained by using a portion of the acquired experimental dataset (i.e., finding the optimal values of its hyperparameters) and by validating it using another portion of the acquired experimental dataset.
[0089] Once the training process is complete, the optimal, validated parameters (such as the weights of the ANN) are loaded into the computer program.
[0090] Preferably, the estimation algorithm is a Mahalanobis Distance-Based Classifier (MDBC).
[0091] Mahalanobis distance is a measure of the distance between a point and a distribution. Unlike Euclidean distance, which treats all dimensions equally, Mahalanobis distance considers both the correlation between variables and the variance within the dataset. Mahalanobis distance is particularly useful for identifying how similar a point is to a given class when considering feature dependencies. This distance metric is especially effective in classification tasks because it normalizes variations across different feature dimensions and provides a scale-independent measure of similarity. For a data point x and a distribution with mean vector x... C The covariance matrix Σ C The category C is represented, and the Mahalanobis distance is calculated as follows:
[0092]
[0093] The estimation algorithm MDBC is used to define Q fixation regions, where each fixation region is associated with a fixation direction formed by pairing the azimuth and elevation angles of the eyeball, and the algorithm is configured to associate the fixation region with each detection vector.
[0094] The estimation algorithm MDBC provides a calibration mode and a prediction mode that can be selected at the start of algorithm execution.
[0095] In calibration mode, the MDBC estimation algorithm includes the following steps:
[0096] - For each fixation region i, where i = 1, ..., Q, collect Z samples, which are the Z detection vectors obtained when the user views the i-th fixation region;
[0097] - For each fixation region i, where i = 1, ..., Q, estimate the mean and covariance of the Z collected samples.
[0098] In the prediction mode, the estimation algorithm MDBC includes the following steps:
[0099] - For each fixation region i, where i = 1, ..., Q, calculate the Mahalanobis distance between the detection vector and the statistical representation (mean and covariance) of the i-th fixation region;
[0100] - Associate the detection vector with the gaze region that has the minimum distance from this detection vector.
[0101] In the estimation algorithm MDBC, each gaze direction represents a category.
[0102] Preferably, at least two infrared radiation sources 210 are configured to emit infrared radiation at corresponding different wavelengths.
[0103] This allows for even more accurate eye tracking because different parts of the eye (such as the fundus) exhibit reflectivity that varies with wavelength. For example, if the scleral reflectivity is constant within a selected IR spectral range, while the fundus reflectivity increases with increasing wavelength, then by selecting two wavelengths λ1 and λ2 (e.g., 850 nm and 950 nm, respectively) in the IR region and comparing the reflected signals R(λ2) / R(λ1), better contrast between different tissues can be obtained compared to the case with a single wavelength.
[0104] Preferably, the processing and control unit 230 is configured to control at least two infrared radiation sources 210 such that the at least two infrared radiation sources simultaneously emit corresponding modulated infrared radiation.
[0105] Preferably, the processing and control unit 230 is configured to control at least two infrared radiation sources 210, such that the at least two infrared radiation sources emit corresponding modulated infrared radiation sequentially. In this case, each light source is turned on at a specific time and in a fixed time sequence, one at a time. Furthermore, the processing and control unit 230 is configured such that its receiving unit for receiving electrical detection signals is time-synchronized with its modulation unit for modulating the infrared radiation of the at least two infrared radiation sources 210. Therefore, the processing and control unit 230 can uniquely identify the source by its time position along the sequence.
[0106] Time-division multiplexing and frequency modulation can be used independently or in combination. Amplitude modulation can be achieved using square waves (OOK, on-off keying) or other periodic waveforms such as sine waves. If a single infrared radiation source is turned on each time (which is constant or modulated at a given frequency, for example to shift the signal into a spectral region characterized by less noise), time-division multiplexing is used to identify individual sources. Conversely, if multiple sources are turned on simultaneously, different modulation frequencies are needed to identify individual sources.
[0107] Using a modulated signal with a non-zero frequency enables better identification of modulated infrared radiation relative to the background, thus allowing for more accurate estimation of the signal-to-noise ratio (SNR). This improvement in SNR is particularly advantageous for distinguishing different parts of the eye affected by LED light, resulting in greater accuracy in estimating the eye's response.
[0108] Another significant advantage of controlling the emission of at least two infrared radiation sources using time-division multiplexing technology is that the intensity of infrared light exposed to the eye is lower compared to FDM-based systems. This makes the eye-tracking system 200 safer and less invasive because it reduces the overall intensity of the infrared light illuminating the eye and lens. In fact, it is well known that prolonged exposure to infrared light can produce nonlinear effects on different parts of the eye; light absorption is not always proportional to the intensity of radiation, and at certain wavelengths, small increases in power can cause localized heating due to phenomena such as saturation or poor heat dissipation. In this context, maintaining the lowest possible exposure to infrared light (as envisioned in the eye-tracking system 200) is essential to avoid interference during prolonged use.
[0109] Preferably, the eye-tracking system 200 may further include a battery 240 associated with a processing and control unit 230, at least one infrared detector 220, and at least two infrared radiation sources 210 to power them.
[0110] The eye-tracking system 200 may also include a wireless communication module 250 associated with the processing and control unit 230, which is configured to communicate wirelessly with an electronic terminal (such as a smartphone).
[0111] The estimation of gaze direction and other eye parameters (such as blink or pupil diameter) depends on the number and location of the infrared detectors. Optimal locations can be obtained through numerical simulation. In particular, the number and location of at least one infrared radiation source 210 and infrared detector 220 can be varied by means of optical simulation in order to maximize the amount of light collected by the infrared detector 220 and the spatial resolution (i.e., the sensitivity of the collected light map to the gaze direction).
[0112] The number of detectors 220 determines not only the calculated angular resolution of the gaze direction, but also the values of power consumption and dissipation. In particular, the maximum number of detectors 220 results in the maximum angular resolution, but also the maximum values of power consumption and dissipation.
[0113] It is necessary to determine the number of detectors, which represents a trade-off between the need for sufficiently high angular resolution and the need for sufficiently low power consumption and dissipation.
[0114] Preferably, the processing and control unit 230 and the battery 240 are installed in two different parts of the frame. However, they can also be installed in the same part of the frame.
[0115] In any case, the interconnection between at least one infrared detector 220, at least two infrared sources 210, processing and control unit 230, and the battery can be achieved by means of wires or flexible printed circuit boards (PCBs).
[0116] In a particular embodiment, battery 240 includes a photovoltaic cell system adapted to recharge battery 240.
[0117] The eye-tracking system 200 can also be incorporated into a head-mounted device. In this case, the head-mounted device includes a frame configured to be worn on the head, at least one lens attached to the frame for positioning in front of at least one eye, and an eye-tracking system 200 associated with the frame, wherein the eye-tracking system and the at least one lens can be manufactured as described above.
[0118] Based on the description provided, the characteristics and relative advantages of the eye-wearing device (the object of this invention) are clear.
[0119] Finally, it is clear that the envisioned eye-wearing device is readily subject to numerous modifications and variations, all of which are within the scope of this invention; furthermore, all details can be replaced by technically equivalent elements. In practice, the materials used and their dimensions can be of any type, depending on the technical requirements.
Claims
1. An eye-wearing device (100), comprising: - Eyeglass frames (110); - An eye-tracking system (200) associated with the frame (110); The eye-tracking system (200) is characterized in that it comprises: - At least two infrared radiation sources (210) are connected to the frame (110) so that when the user wears the eyewear (100), the eyewear faces the corresponding eye to be tracked and emits infrared radiation toward the eye; - At least one infrared detector (220) is connected to the frame (110), the at least one infrared detector (220) being configured to detect optical signals reflected by the eye to be tracked and generate a corresponding electrical detection signal; - A processing and control unit (230) associated with the at least two infrared radiation sources (210) and the at least one infrared detector (220), the processing and control unit (230) being configured to control each of the at least two infrared radiation sources (210) to emit infrared radiation modulated with a corresponding modulation signal, receive the electrical detection signal, and estimate the gaze direction of the eye to be tracked based on the received electrical detection signal.
2. The eye-wearing device (100) according to claim 1, wherein, The modulated signals have different frequencies from each other.
3. The eye-wearing device (100) according to claim 1 or 2, wherein, The modulated signals have different shapes from each other.
4. The eye-wearing device (100) according to one or more of the preceding claims, wherein, The modulated signals are substantially orthogonal to each other.
5. An eye-wearing device (100) according to one or more of the preceding claims, wherein, The at least two infrared radiation sources (210) are configured to emit infrared radiation at corresponding different wavelengths.
6. An eye-wearing device (100) according to one or more of the preceding claims, wherein, The processing and control unit (230) is configured to control the at least two infrared radiation sources (210) such that the at least two infrared radiation sources simultaneously emit corresponding modulated infrared radiation.
7. An eye-wearing device (100) according to one or more of the preceding claims, wherein, The processing and control unit (230) is configured to control the at least two infrared radiation sources (210) such that the at least two infrared radiation sources emit corresponding modulated infrared radiation in sequence.
8. An eye-wearing device (100) according to one or more of the preceding claims, comprising a battery (240) associated with the processing and control unit (230), the at least one infrared detector (220) and the at least two infrared radiation sources (210) for powering them.
9. An eye-wearing device (100) according to one or more of the preceding claims, wherein, The processing and control unit (230) is configured to filter the received electrical detection signal with at least two noise filters, one noise filter for each infrared radiation source, wherein the noise filters are configured to reduce the noise contribution in the received electrical detection signal.
10. The eye-wearing device (100) according to claim 9, wherein, The noise filter is configured to minimize the noise contribution in the received electrical detection signal, and the noise filter has a time transfer function equal to the mirror time function of the modulation signal.
11. An eye-wearing device (100) according to one or more of the preceding claims, wherein, The processing and control unit (230) is configured to implement an estimation algorithm to estimate the gaze direction of the eye to be tracked based on the received electrical detection signal, the estimation algorithm being a Mahalanobis distance-based classifier (MDBC).
12. A head-mounted device, comprising: - A frame configured to be worn on the head and at least one lens, the at least one lens being attached to the frame to be positioned in front of at least one of the user's eyes; - An eye-tracking system (200) associated with the frame (110); The eye-tracking system (200) is characterized in that it comprises: - At least two infrared radiation sources (210) are connected to the frame (110) so that when the user wears the eyewear (100), the eyewear faces the corresponding eye to be tracked and emits infrared radiation toward the eye; - At least one infrared detector (220) is connected to the frame (110), the at least one infrared detector (220) being configured to detect optical signals reflected by the eye to be tracked and generate a corresponding electrical detection signal; - A processing and control unit (230) associated with the at least two infrared radiation sources (210) and the at least one infrared detector (220), the processing and control unit (230) being configured to control each of the at least two infrared radiation sources (210) to emit infrared radiation modulated with a corresponding modulation signal, receive the electrical detection signal, and estimate the gaze direction of the eye to be tracked based on the received electrical detection signal.
Citation Information
Patent Citations
Eye tracking system
US20180113508A1
Infrared eye-tracking in high ambient light conditions
US20190384388A1
Eye tracking system and method
US5345281A
